Smart People want Smart Questions not Dumb Answers.

Thinking of…Building a Digital Operating Model with the Microsoft Cloud Adoption Framework? Ask the Smart Questions
So this book deliberately does not provide generic answers. Instead it is a
comprehensive set of questions that will enable you to find the answers that are
right for you. Enabling you to make the critical decisions.

“Judge a man by his questions rather than by his answers” VOLTAIRE

We live in a time of unprecedented change. Digital and broader technology

innovation are reshaping our world all around us. The introduction of the cloud
and on-demand access to computing power have turbo charged this already fast
rate of innovation and change.
Embracing cloud in your organisation is more than just moving servers from on
premises to someone else’s datacentre. That part is the easy part. Fully embracing
cloud necessitates taking a long, hard look at your organisation. What does digital
mean to you? How do you go about digitally transforming your organisation?
This conversation extends far beyond the remit of your IT department. As leaders
within your business, simply delegating the problem to IT is not the answer.
Technology in itself is not the answer. The answer falls exactly between the
traditional realms of business and IT.
The primary currency of a business operating model is agility. The primary focus of
an IT operating model is control. The key to digital transformation is to seamlessly
blend these magnetically opposite operating models. Only when these teams and
these concepts truly combine can you be successful in this digital age. This book
argues the need to establish a cloud operating model as this unique bridge between
the two. A seamless combination of business and IT operating models.

“Building a Microsoft Cloud Operating Model is a must read for leaders looking to
understand how the rules of the game have changed, and importantly how to unlock
the value that comes with the right model, great technologies and engaged people.
I love the fact it’s practical and serves as a useful guide for those driving change and
innovation in their business”
Secure
CLARE BARCLAY, CHIEF OPERATING OFFICER, MICROSOFT UK Compliant
Agile
Thinking of…
Building a Digital Operating Model with the
Microsoft Cloud Adoption Framework for Azure?
Ask the Smart Questions
www.smart-questions.com By Dan Scarfe, Sean Morris,
Frank Bennett and Ray Bricknell
..MS2019..
Thinking of...

Building a Digital Operating

Model with Microsoft Cloud
Adoption Framework for
Azure?

Ask the Smart Questions

By Dan Scarfe, Sean Morris, Frank

Bennett and Ray Bricknell

..MS2019..
Copyright © 2019 Dan Scarfe, Frank Bennett, Ray Bricknell
and Sean Morris
First Published in 2019 by 1visionOT Pty Ltd trading as Smart Questions, Suite 3, 596
North Road, Ormond, VIC 3204, Australia
Web: www.smart-questions.com (including ordering of printed and electronic copies, extended
book information, community contributions and details on charity donations)
Email: [email protected] (for customer services, bulk order enquiries, reproduction
requests et al)

The right of Dan Scarfe, Frank Bennett, Ray Bricknell and Sean Morris to be identified as
the authors of this book has been asserted in accordance with the Copyright, Designs and
Patents Act 1998. All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system or transmitted, in any form or by any means, electronic,
mechanical, recording or otherwise, in any part of the world, without the prior permission
of the publisher. Requests for permission should be sent to the publisher at info@smart-
questions.com

Designations used by entities to distinguish their products are often claimed as trade
marks. All brand names and product names used in this book are trade names, service
marks, trade marks or registered trademarks of their respective owners.

The authors and publisher have taken care in preparation of this book, but to the extent
permitted by applicable laws make no express or implied warranty of any kind and assume
no responsibility for errors or omissions. The contents of this book are general in nature
and do not constitute legal or other professional advice. You should always seek specific
professional advice relevant to your circumstances. Where specific costs or prices have
been given, these represent our best ‘standard case’ estimates at the time of writing. They
are intended to provide a rough indication only and should not be relied on in any way.
To the extent permitted by law, no liability is assumed for direct, incidental or
consequential loss or damage suffered or incurred in connection with or arising out of the
use of the information contained herein.

A catalogue record for this book is available from the British Library.

ISBN 978-1-907453-27-4

SQ-23-198-001-001

..MS2019..
Smart QuestionsTM Philosophy
Smart Questions is built on 3 key pillars, which set it apart from other
publishers:
1. Smart people want Smart Questions not Dumb Answers
2. Domain experts are often excluded from authorship, so we are making
writing a book simple and painless
3. The community has a great deal to contribute to enhance the content

www.smart-questions.com

..MS2019..
ii

..MS2019..
Reviews
Clare Barclay
Chief Operating Officer, Microsoft UK

“Building a Digital Operating Model with the Microsoft Cloud Adoption

Framework for Azure” is a must read for leaders looking to understand how
the rules of the game have changed, and importantly how to unlock the value
that comes with the right model, great technologies and engaged people.
I love the fact it’s practical and serves as a useful guide for those driving change
and innovation in their business.”

Evelyn Padrino
Azure Marketing Director, Microsoft

Adopting ‘the cloud’ carries core changes for an organization, that go beyond the
obvious technology aspects, incorporating business and process components
needed for a successful digital transformation.
“Building a Digital Operating Model with the Microsoft Cloud Adoption
Framework for Azure” provides a balanced and practical approach to operate
in this new ‘cloud era’. With its easy, and almost entertaining, prose, the book
guides the reader through this transformation. Presenting insightful questions
and ‘real-life’ references, it triggers internal conversations across the business and
technology sides of the organization to identify and execute the right path for
cloud adoption.

iii

..MS2019..
Authors

Dan Scarfe
Dan is a passionate technologist and
loves envisioning technology solutions
that positively impact our culture and
society.
New Signature is in the business of
helping enterprise organizations
harness the power of the Microsoft
Cloud.
As EVP Global Solutions, Dan heads up the sales and marketing
GTM vision conceiving new products and services which can help
organizations digitally differentiate, working closely with the sales
and delivery teams.
Dan has been deeply engaged with Microsoft's Cloud since its
inception in 2008. He was advising Microsoft on Azure when it
was still called Red Dog and has seen the platform evolve
substantially over that time. He presents on this topic around the
world and sits on the Azure Partner Advisory Council.
Dan has been involved in many start-ups over the years, some
successful, some not, and does his bit through passing on lessons
learned as a mentor within Microsoft for Start-ups.
Dan has authored two other books in the Smart Questions series.

iv

..MS2019..
 Frank Bennett FRSA
Frank has a long career in Information
Technology from the distant age of the
mainframe and every evolution
preceding the Cloud era.
He is Smart Questions’ most prolific
author with eight published titles. As well
as writing about Cloud computing he is a
practitioner and recently co-founded a
Software as a Service business to assist
charities and not-for-profit organizations with the improvement of
their governance.
He is a mentor to Microsoft ScaleUp London. An independent
expert on the European Union’s Horizon 2020 CloudWatch2
program. He is attributed as the inventor of Market Readiness
Levels and their conjoining with Technology Readiness Levels. In
2018 he gifted the invention to Oxford University Innovations for
use by Research & Innovation projects in academia.
Now with a portfolio career he holds the Financial Times Non-
Executive Director Diploma. He is qualified to advise
organizations on the General Data Protection Regulation (GDPR)
following completion of the GDPR Transition Programme at the
world-renowned Henley Business School.
Frank is Deputy Chair of the UK Cloud Industry Forum. A Fellow
of The RSA and Member of EY Independent Director
Programme.

..MS2019..
 Ray Bricknell
With 37 years of international IT
experience, Australian entrepreneur
Ray Bricknell has been actively
facilitating and commenting on the
relationship between the mid-tier UK
Financial Services sector and the UK
Cloud Vendor community since 2010.
Ray’s firm Behind Every Cloud
developed the award winning and industry accredited Clover
(Cloud Vendor Ratings) index. He provides advisory services to a
diverse range of large and small clients from the Asset
Management/Investment Management/ Hedge, Private Equity,
Retail, Investment and Private Banking and Insurance. Formerly
the CTO of an $8Bn listed Hedge Fund. Ray is currently Chair of
the Cloud Industry Forum Financial Services Special Interest
Group.

Sean Morris
Sean has been involved in the IT
industry since the early 90s and has sat
on both sides of the consultancy world
as an in-house IT staffer and a
consultant working for large systems
integrators.
Sean has a background in professional
services from working as part of an IT
team for a top London-based law firm and in Telco working in
Australia. Sean specializes in infrastructure services particularly
around hosting and service management.
Sean runs a team of Cloud advisors and pre-sales consultants for
New Signature UK, pretending he works for Dan.
Sean is passionate about the business of IT and the role IT can play
in contributing to the success of organizations.

vi

..MS2019..
Table of Contents

Part 1 – Business
'What the leadership team need to know'

1 The Digital Age .................................................................................... 1 

2 The Digital Wave ................................................................................. 7 
3 The Way Ahead .................................................................................. 17 
4 Digital Countdown – The Race is on.............................................. 25 
5 Agile – small word quick work......................................................... 37 
6 The Business Questions .................................................................... 45

Part 2 – Technical
'What the technical team need to know'

7 A new operating model ..................................................................... 79 

8 Strategy and service providers .......................................................... 83 
9 Procurement and financial governance........................................... 97 
10 Service management ...................................................................... 105 
11 Access control, security and policy.............................................. 123 
12 Monitoring, management and automation ................................. 139 
13 DevOps and application development ....................................... 147 
14 The Technical Questions .............................................................. 165 
15 Funny you should say that ............................................................ 203 
16 Final Word ...................................................................................... 211 

vii

..MS2019..
Acknowledgements
We’d like to offer huge thanks to everyone that has helped and
supported us as we wrote this book. The list is too long to thank
each person, but particular thanks to:
 Mark Smith for championing this initiative inside Microsoft
and encouraging us to do this.
 Stelios Zarras for your help and support along the way.
 Brian Blanchard and the rest of the Cloud Adoption
Framework for Azure team at Microsoft for giving us the
content to support this book.
 Evelyn Padrino, Pratibha Sood and Sonia Yu for supporting
the new edition of this book.
 John Kendrick for sharing his valuable insights in the first story
at the end of the book.
 Pete Gatt from Servian who contributed a huge amount to the
thought around some of the concepts we describe, along with
helping to write a good chunk of the book and sharing the
second story at the end of the book.
 New Signature for giving us the time to write this book.
 Jane Scarfe for proof reading.
 Lara Scarfe for letting Dan spend endless evenings and
weekends writing.

viii

..MS2019..
Foreword
Mark Smith, General Manager,
Microsoft Solutions
We live in a time of unprecedented
change. Digital and broader
technology innovation are reshaping
our world all around us.
The introduction of the Cloud and on-
demand access to computing power
far in excess of anything available to
organizations before have turbo-charged this already fast rate of
innovation and change.
In recent years, the Cloud has evolved from a set of technologies
that augment mission-critical platforms, to become that mission-
critical platform. It is no longer something used for lower impact
workloads. It is becoming the central platform organizations are
leveraging as the lynchpin of their digital transformation strategies.
As we hurtle along on this journey of change, it’s sometimes
difficult to give ourselves the time to take stock. To take the time
to look back on what we have achieved. To assess what has worked
and what requires more effort. To learn from others about their
successes and failures. To understand and evaluate what good
looks like.
This book seeks to do just this.
Embracing Cloud in your organization is more than just moving
servers from on premises to someone else’s datacenter. That part is
the easy part. Fully embracing Cloud necessitates taking a long,
hard look at your organization. What does digital mean to you?
How do you go about digitally transforming your organization?
What aspects of technology can you embrace to allow you to not
only survive but also thrive in this brave new world?
This conversation extends far beyond the remit of your IT
department. As leaders within your business, simply delegating the
problem to IT is not the answer. Technology in itself is not the
answer. The answer falls exactly between the traditional realms of

ix

..MS2019..
business and IT. To truly digitally transform your organization your
business leaders need to understand the role technology can play
and imagine new technology-powered products and services. At the
same time, your technology leaders need to far more closely align
themselves with the business. It is these teams and these individuals
who are uniquely placed to transform these ideas into reality.
The key to digital transformation is to seamlessly blend what we
might describe as a business operating model with our traditional
IT operating model. Only when these teams and these concepts
truly combine can we be successful in this brave new world.
This book argues the need to establish a Digital Operating Model
as this unique bridge between these two separate worlds. A
seamless combination of business and IT operating models.
A combination of these two completely separate worlds is,
however, fraught with difficulty.
The primary currency of a business operating model is agility.
Business owners and business group leaders thirst this agility. They
are often perplexed as to why things are perceived as being so
difficult and time consuming. They are focused on customers and
delivering them the capabilities they desire. Everything that gets in
the way of this is unnecessary complexity and roadblocks.
The primary focus of an IT operating model is control. IT leaders
wake at night worrying about availability, security and a raft of
factors completely alien to the business. IT often shies away from
innovation and change as it is often in direct competition with their
driving goals.
This mismatch is one of the primary contributors to the “shadow
IT” phenomenon of recent years. Business leaders demanding a
level of agility not being adequately delivered by the IT department
simply pull out their corporate credit cards and procure what they
need, there and then.
Uncontrolled and unabated procurement of IT solutions is not a
good long-term strategy. Control is still a necessity within larger,
structured organizations.
A Digital Operating Model as described within this book seeks to
achieve the best of both worlds. The agility demanded by the
business with the control needed by the IT teams.
x

..MS2019..
Preface
Brian Blanchard, Sr. Director,
Microsoft Cloud Adoption
Framework for Azure
We are living in the most innovative
era in human history. Developers are
delivering global user experiences,
creating and deploying applications
faster and at greater scale than ever
before. The value of data continues to
skyrocket. Those who harness data and AI are changing entire
markets. As markets change and expand, infrastructure engineers
are making businesses more agile and better equipped to capitalize
on change. Together, infrastructure, apps, and data are reshaping
businesses and driving innovation.
These changes, often referred to as the “Digital Transformation”
movement, are happening in every industry and every region of the
globe. The Cloud is a catalyst for this change, but architects,
engineers, and business leaders are the real source of this
innovation. They provide the human potential that is changing the
business world.
Harnessing this human potential requires something Microsoft
refers to as “Technical Intensity”. Developing technical intensity
requires a combination of technical capability, technical adoption,
and trust. This concept hints to the correlation between cloud
technologies and rapid growth in innovation.
For decades, technical capability was constrained by a company’s
ability to acquire hardware. The cloud has removed most of the
capital expense blockers to innovation, paving the way for technical
intensity and the ensuing innovation.
Companies that have most successfully driven innovative business
change have matched technical capability with an operating model
that encourages adoption of technology across the business and
customer base (Tech adoption). More importantly, those
companies have done so with sound governance and operational
management processes to provide safe guardrails (Trust).

xi

..MS2019..
This book builds on Microsoft Cloud Adoption Framework for
Azure to guide customers as they create their own operating model
for the digital era. This book and the supporting framework guide
readers as they prepare the business, people, and processes to
deliver tech intensity and lasting innovation.
Beyond the book: Microsoft works with a broad range of
companies from start-ups to the world’s largest enterprises. For
years we supported companies as they delivered on cloud adoption
and built their own operating models. Along the way, we’ve asked
thousands of employees, partners, and customers about their
experiences and best practices.
The Cloud Adoption Framework is a collaborative effort across
Microsoft to demonstrate agile principles and a growth mindset.
Together, we’ve learned from what works (and what doesn’t).
Together we’ve documented business, culture, and technology
lessons that make Cloud adoption easier. The Cloud Adoption
Framework is the output of those lessons, creating a collection of
tools and documentation to guide companies through the iterative
phases of the cloud adoption lifecycle.

Making it real: Throughout this book, the authors link to various

assets in the framework to guide you through your Cloud adoption
journey. Those links and the authors’ experiences come together in
this book to describe a Digital Operating Model. The new set of
rules. The new common environment where everyone can focus
together on how to make our companies digital businesses. As you
build out your company’s Digital Operating Model, build technical
intensity, or just deliver on a Cloud adoption project, we hope you
will join us in expanding the Cloud Adoption Framework by
sharing your experiences and feedback!

xii

..MS2019..
Introducing the Microsoft Cloud
Adoption Framework for Azure
The world around us is changing at a rate none of us has observed
before. The proliferation of access to scalable, pay as you grow,
computing resources has changed the rules on creating new
technology-powered experiences for customers. The rate of
innovation continues to speed up every day, every week and every
month.
But what should you do if this sounds like something you want to
do? What does it mean to become a digital company? How can you
leverage technology to differentiate your services in market? How
can you prepare your people, technologies, and processes for this
digital transformation?
An adoption framework is the lingua franca of all three major
public Cloud providers in answer to these questions. Amazon and
Google also both have their own adoption frameworks. Microsoft’s
adoption framework is the newest entrant and the authors would
suggest the most comprehensive
As Cloud computing has become mainstream it raises many
questions and demands a framework for the resulting
conversations. How do we create a digital strategy? What does that
mean?
It is ultimately the myriad of decisions about how an organization’s
resources – technology and people – are organized to deliver
business outcomes. Those business outcomes are defined and
designed by decisions of people. And so, the Cloud Adoption
Framework is a framework for people to collaborate and put
technology to work.
It is unlikely that any one person will be an expert in all aspects of
the Cloud Adoption Framework. Its application will involve the
collaboration of a multi-disciplinary team. The purpose of this
book is to provide the ‘go to’ resource for that team.

xiii

..MS2019..
The Cloud Adoption Framework for Azure is divided into 8
sections:

It starts with Strategy. What does it mean to be a digital company?

Why is it that you are embarking on this journey?
It then helps you to Plan. What does your digital estate look like
today? What is the make-up of your team today? What’s the relative
appetite for this change across the organization?
The next stage is to get Ready. You’ll need to start to build out
your base Azure environment. Think about how to structure it.
Iterate and continue to add additional capabilities and services.
In parallel, it’s vital you consider how you will Govern your
environment. What are the key dimensions of governance? What
controls will I need to put in place?
Once your environment is built and governed, you’ll need to think
about how you Manage it. What tools are you going to use? How
are you going to operate it? How will Cloud plug in to your existing
IT service management organization?
The final piece in the puzzle is to start to Adopt it. To start you
might just Migrate workloads from on premises. In the future, you
might be creating new digital experiences to help you Innovate and
differentiate within your market.
A key differentiator in the digital age is developing a business
culture that is agile. It is a culture that anticipates and feels at home
with constant change. Those that have got to grips with
implementing agile, report that it resulted in a shake-up of
organizational structures. Mindful of that, it is recommended you
read Chapter 5, so you are ready for the conversation.

xiv

..MS2019..
The Digital Operating Model is where everyone comes together –
Business teams, Developers, IT Operations. It is where discussions
about ‘digital at the core’ of delivering business outcomes are
turned into actionable plans. For some that will be the creation of a
digital business. For others it may be a toe in the water with Cloud
computing. Those organizations pursuing a digital transformation
strategy will find the Cloud Adoption Framework is highly suited
to the orchestration of that and we hope this book provides some
context as you embark on your journey.

xv

..MS2019..
Who should read this book?
This book is presented in two parts.

Part 1 Business is for the non-technical reader, those concerned

with shaping the organization.
Part 2 Technical is for the learned technical reader, those
concerned with solving how technology supports the organization.
We encourage readers to read both parts so when non-technical
and technical audiences come together there is a mutual
understanding of what is on the mind of the other.
This book is about the creation of a Digital Operating Model. What
we mean by that is how your organization consumes the resources
of public Cloud-based services to compete in this digital age. In
chapter 15 Funny You Should Say That, we have invited people
who have been through the transformation to write up their
experiences.

People like you and me

This book is aimed squarely at those who want to effect a change
in the way IT services, and specifically Cloud services, are
consumed and delivered to an organization. Those who recognize
the vital role IT plays in a world that is increasingly technology-
driven in what is commonly referred to as the ‘digital age’. Those
that want to become change agents. People like you and us.
This book is intended to be a catalyst for action aimed at a range of
players inside and outside your organization. Here are just a few of
those players, and why it is relevant to them.

The Board
The board has many things to consider. Every day, technology gets
higher and higher on the priority list. With the rate of technological
innovation and the pervasive use of Cloud computing it is a
juggling act to not get left behind – never mind get ahead.

xvi

..MS2019..
This book describes a model for how a business will make
decisions to deal with ever-increasing reliance on technology.
Those decisions are complex as they can conflict with the way IT is
delivered today and previous investments you may have made. It’s
difficult to balance this with taking advantage of the innovation
served up in the Cloud. Don't be fearful that you will drown in
technical jargon. Part 1 avoids this and sets questions you will want
to ask of yourself and others:
As you look forward, what technological capability will the business
require? What skills and experience will you need in order to put
innovation to work? Who do you partner with for knowledge and
support?
Boards will be aware of the techno-political debate that places a
new responsibility on organizations regarding the processing of
data. A legal responsibility and accountability now accompany the
decision to invest in technology such as Artificial Intelligence for
the processing of data.
The connectedness and ease with which data can transmitted in the
digital age, while creating new opportunities, also comes with new
responsibilities. Data privacy has jumped up the agenda of
politicians, regulators and consumers. In some jurisdictions, such as
the European Union, the consequences of non-compliance are
severe. Others are following. The California Consumer Privacy Act
is expected to come into force on Jan. 1, 2020. This is a governance
matter for the board’s attention.

The CxO team

You might think that technology doesn’t apply to you. It’s an
implementation detail within your organization. Something for
somebody else to worry about. This is not true anymore. The
discussion is becoming ever more complex as business and
technology are interwoven and are core to strategy and execution in
the business. Management has the ultimate responsibility for
business outcomes and today’s leaders must operate in the style of
leaner ways of thinking and operating, as revealed in this book.
A measure of a successful company is the way it harnesses
technology. That is both an offensive and defensive strategy. Our
world has a long list of established businesses which have gone
xvii

..MS2019..
from glory to bust. Then follows the 'Oh poor them' analysis of
what happened. We all know who they are – we used to shop there;
they might have been a supplier.
So, if it is agreed that technology is now make or break then this
book is for you. Engage the Board and Line of Business Managers
in a discussion about what the business must be good at. That
could be customer service, optimization of the supply chain,
supporting a mobile workforce. These things describe how the
business operates. This book serves up a discussion about how that
evolves with technology and harnessing the power and innovation
of Cloud computing.

CIO / CTO / CDO

You as the CIO, CTO or CDO are key architects of the
transformation we describe in this book, and your role will change
fundamentally. You’re going to move from expert execution to
expert orchestration and transformation. The execution of the
Digital Operating Model is in your hands.
The Cloud Adoption Framework is your roadmap. You may want
to align with partners who bring experience to your discussions,
plans and execution.

Line of Business Manager

There are no better people to tell you about the demands of the
business than those at the sharp end. You may hear that your
competitors are delivering new products and services more quickly.
Or it may be that they are making it easier and more convenient for
customers to conduct their day-to-day business.
The availability of Cloud computing has created new opportunities
for Line of Business managers to respond quickly. It is they who
can equip their teams with new ways to deliver productivity and to
work smart. That has served its purpose well and is now part of the
big picture of how the business will create an agile posture to the
deployment of technology.

xviii

..MS2019..
How to use this book
This book is intended to be the catalyst for action. We hope that
the ideas and examples inspire you to act. So, do whatever you
need to do to make this book useful. Use Post-it notes, write on it,
rip it apart, or read it quickly in one sitting. Whatever works for
you. We hope this becomes your most dog-eared book.

Smart Questions
At the end of each part you will see a table of questions. Not all the
questions will necessarily be new or insightful. The value you get
from the information will clearly vary. It depends on your job role
and previous experience. We call this the 3Rs.
Some of the questions will be in areas where you know the answers already,
so the questions will Reinforce them in your mind.
You may have forgotten some aspects of the subject, so the questions will
Remind you.
Other questions may Reveal new insights to you that you’ve never considered
before.

We trust that you will find real insights. There may be some “aha”
moments. In this context, probably the most critical role of the
Smart Questions is to reveal risks that you might not have
considered. On the flip side they should also open your thinking to
opportunities that hadn’t yet occurred to you. Balancing the
opportunities and the risks, and then agreeing what is realistically
achievable is the key to formulating an effective strategy.
The questions could be used in your internal operational meetings
to inform the debate. Alternatively, they could shape the discussion
you have with your IT vendors and their partners.

xix

..MS2019..
xx

..MS2019..
Part 1 Business

Thinking of

Building a Digital Operating

Model with the Microsoft Cloud
Adoption Framework for Azure?

Ask the Smart Questions

'What the leadership team need to know'

xxi

..MS2019..
xxii

..MS2019..
 The Digital Age

Chapter

1
1The Digital Age
Someone is sitting in the shade today because someone planted a tree a long time
ago
Warren Buffett (Businessman, investor and philanthropist, 1930-)

A new dawn

A ROUND the world the disruptive promises of the

infamous dot com bubble of the late nineties and early
noughties are finally being realized. The current generation
of IT that had Cloud enter everyday vocabulary has
businesses embracing digital transformation and digital-enabled
business methods. As a result, a new requirement has emerged for
a different kind of Operating Model – a Digital Operating Model
(DOM). As we will explain, the Digital Operating Model is set to
become the new lingua franca of business.
The work behind the Digital Operating Model where multi-skilled
teams come together and collaborate is codified as the Microsoft
Cloud Adoption Framework for Azure. Do not be confused, they
co-exist. It is likely the CxO suite will be more comfortable to talk
about the Digital Operating Model - it has a business ring to it. The
teams tasked with delivering the Digital Operating Model will
reference the Cloud Adoption Framework. More on this later.
The digital agenda is commanding CxO attention worldwide. Its
outcomes will set the scene for high-speed business change over
the next decade. This will spell opportunity for some and hazard
for others. With constant change that appears to be accelerating
new demands are put on organizations to respond.
1

..MS2019..
The Digital Age

The big shift is that organizations must develop a culture and

competency to be agile. For some, particularly those bureaucratic
organizations, that is a big ask. For this reason, we explore the
balance between agility and control as a key theme in this book. We
advocate the necessity of (some) sacrifice of control in order to
enable the agility now demanded by those operating in a digital
world. Moving the dial toward agility will please business teams.
However, it will challenge IT teams accustomed to control as a
doctrine.

A new day, a new way

One of the unforeseen side effects of the adoption of Cloud is
‘shadow IT’. This term describes infrastructure and applications
purchased or sometimes developed by the business without the
engagement of the IT department. Why did that happen? Quite
simply the Cloud served up new ways for business teams to get
things done fast without the need to be (or engage) technologists.
Shadow IT supports line of business to move fast and gives them
the agility to do what they need to do. It is also accounted for
within delegated budgets and treated as an expense rather than an
asset. Some say this has resulted in a loss of control of all-up IT
expenditure. They claim that governance, standardization,
economies of scale and overall strategic coordination suffer. Ho
hum! Things are getting done and that’s what matters most. Right?

New horizons
The prevailing best practice for centralized control of IT/business
oversight is the formality of the IT Operating Model (ITOM). This
is an aid to communication, coordination and control. Intermediary
roles and translation functions like IT project managers and
business analysts exist to define and agree the following:
 The requirements and priorities of the business
 The objectives and constraints before work starts. The
budget and forecast future spend
 Formal plans and reporting for transition and migration
 Structured processes to align the goals and work, such as
Waterfall and PRINCE2 ™ etc. Check out the Appendix
for a definition.
2

..MS2019..
 The Digital Age

The ITOM model can frustrate business teams’ projects. There is

always a queue of corporate priorities that are seemingly perpetually
delayed. If they want results now, many have been tempted by the
shadow IT option.
Shadow IT costs continue to skyrocket. In some cases, this can
create data silos and system duplication. Yet still, it is no barrier to
the insatiable appetite for Cloud – as we have witnessed through its
meteoric growth. This has not gone unnoticed by the senior
executive team and the Board. They want oversight as IT is a
strategic resource. There is the hint as to why we wrote this book -
to be learned in ways to meet the challenge of aligning an organization’s
resources to grasp new opportunities of the digital age.
 As Cloud went mainstream a host of questions were
initially raised.
 Can we really trust the safety of our data to the Cloud?
 Will we get locked into a vendor with no get-out route?
 What IT stays on-premises and what goes to the Cloud?
 Are the economics of the Cloud inferior or superior?
 Will the Cloud take away my job?
Depending on your point of view, these questions have been
resolved. In any event, we have moved on.
The new questions are about how to harness the Cloud and not get
left behind. For many that is in progress and adoption is growing
rapidly - look at the success of Microsoft 365. Next up is how to
take advantage of the continuous innovation in the Cloud. That is
one important aspect of the Cloud Adoption Framework - putting
innovation to work. The work ahead is to create an agile
organization that puts it arms around the Cloud and each
organization will have its own ideas about that. That will need
context and we describe that as ‘delivering your Digital Operating
Model using a Cloud Adoption Framework’.

Disruption ahead?
Right now, business leaders are wondering what the future is for
their own business models as evidenced below.

..MS2019..
The Digital Age

Figure 1 - Disruption by industry

Cloud technology enabled a wave of disruption and some sectors

of the economy have had their business model disintermediated or
destroyed. It all seemed to happen in the blink of an eye.
Some sectors (such as retail) are probably forecasting less
disruption because they think it has already substantially occurred.
Maybe they are right, maybe wrong. Others such as Financial
Services (especially Banking and Insurance) appear to be bracing
themselves for disruption.
Whatever lies ahead is going to result in change and there will be
barriers. Of course, you can always bury your head in the sand but
that could result in a bad ending. Alternatively read this book to
kick start conversations and then get going.

It is a digital conversation
Many conversations are taking place, like this one:
Q. Do we have a digital strategy?
A. Do we need one?
Q. What do you mean by digital?
A. Well, it is something I am reading a lot about and it could be important to
our future.
Q. What is it exactly?
A. Er, I am not sure. Who can we ask?
4

..MS2019..
 The Digital Age

There is no recognized definition for a digital strategy; it is

whatever is right and appropriate for your business. Many people
are involved in the conversation and in consequence it can get
ragged. You have probably guessed what is coming next. You need
something to organize the conversation around. That something is
the Digital Operating Model and the Cloud Adoption Framework.
There, we said it again!
It is widely understood that a digital strategy is not an IT strategy.
Do not focus exclusively on technology developments – although
technology components will be major elements in the delivery of
the benefits of your Digital Operating Model.
The challenge for many is to bring together the differing ideologies
of the business and IT teams. Business teams seek agility to
respond quickly to the external environment. IT teams do not
always get that. IT teams seek control to cope with the complexity
of IT environments. Business teams do not always get that. Both
need to identify with something they can agree to coalesce around.

The new reality

The IT team is coming to terms with the fact that the control of
their carefully managed IT platforms is too slow. The external
facing business teams need agility for their ‘digital at the core’
competitive stance. The business is demanding speed of change,
and the IT team is on notice to work with them to deliver that.
The writing is on the wall. ‘We need a way to do new stuff together
faster than before. We need it now!’
So, read on to learn more about why a Digital Operating Model is
at the core of every successful 21st century business.

..MS2019..
The Digital Age

..MS2019..
 The Digital Wave

Chapter

2
2The Digital Wave
The world as we have created it is a process of our thinking. It cannot be
changed without changing our thinking
Albert Einstein (Theoretical Physicist, 1879-1955)

B USINESSES generally succeed because they continually

evolve, although not always at great speed. Increasingly that
is enabled by digital technology and talked about as
innovation and transformation.
That occurs in a couple of ways:
 Take an existing business model and execute it in a better
way than competitors. Deliver incremental improvements
ideally at low risk and at low cost. Experiment and accept
failure or hopefully succeed in your outcomes.
 Devise an entirely different approach to meeting customers'
needs and disrupt the status quo. Figure out how to serve
the under-served customer segment(s) with a
complementary economic model and win big.
A business that delivers broadly the same as others and in the same
way, might differentiate on service quality. It might offer improved
speed or accuracy of delivery. It might have intangible metrics such
as ‘easy to do business with’ or a strong brand image. The
difference might often just be doing the same at a lower price. For
established businesses a step change in the business model is
typically harder to achieve than was bargained for. The process of
delivering improvements is easier, but where that relies on
associated IT upgrade(s) then it often happens rather too slowly.
7

..MS2019..
The Digital Wave

That is the soft underbelly that the disrupter attacks as they move
fast and often in stealth mode.

The D word
Talk about disruption conflates
with new business models. They
use technology to completely
change how a product or service is served up to the consumer.
Three very well-known examples are Uber, Netflix and Airbnb.
These three companies entirely changed the way their respective
services were taken to market. The end customers’ consumption
was largely unchanged - it was the buy and fulfilment model that
changed. These disrupters deliberately and intentionally attacked
the established taxi, video rental and accommodation booking
markets. Note that the accommodation booking market had
previously been disrupted so the lesson there is – don’t be
complacent.
Whereas it was once the case that online was only for the brave
owing to security fears (often hyped by antagonists), today there are
very few businesses who don’t transact online. You must also be
exceptional, because other choices are just a click away.
This is today's agenda, as Microsoft report in its Disrupt yourself or
risk being disrupted: Competing in 2020 1
Will you be disrupted?
Business leaders know their industries are ripe for transformation and are
eager to bring the benefits of technology to their business. In fact, in a new
study by Harvard Business Review ‘Competing in 2020: Winners and Losers in
the Digital Economy 2, 80% of the 783 respondents believe their industry will
be disrupted by digital trends. Most of those (84%) said their industry has
either passed the inflection point of disruption or will pass it by 2020.
Digital leaders are doing today the things they need to do to be successful in
2020. Companies that form their strategies now, shift resources to new digital
initiatives, and redesign their organization and culture will have a distinct
advantage. Micro revolutions occur typically every 12-18 months, so
companies must be in a continual state of transformation.

1 https://enterprise.microsoft.com/en-ca/articles/industries/microsoft-services/disrupt-yourself-
or-risk-being-disrupted-competing-in-2020/
2 https://hbr.org/sponsored/2017/04/competing-in-2020-winners-and-losers-in-the-digital-

economy
8

..MS2019..
 The Digital Wave

History is littered with the graves of businesses which, when facing

a significant trend, made the choice to do nothing. They took the
wait-and-see option. They either considered they were too big to
fail or did not react fast enough. It is wise not be complacent.

In 1965, the average tenure of companies on the S&P 500 was 33 years. By
1990, it was 20 years. It is forecast to shrink to 14 years by 2026. The FTSE
100 (London Stock Exchange) was launched in January 1984 and some 35
years later just 27 of its original members remain listed.

It is equally unwise to trivialize the decision. It is simplistic to

assume or be bossed into blindly accepting that all businesses must
disrupt themselves. Let’s look at some very useful myth-busting
content along these lines from the MIT Sloan Management Review 3
(Feb 06, 2017).
Myth #1: Every company should digitally transform.
Reality: Not every company, process, or business model requires
digital transformation.
Myth #2: Digital transformation leverages emerging or
disruptive technologies.
Reality: Most short-term transformational impact comes from
conventional operational and strategic technology – not from
emerging or so-called disruptive technology. That is to say,
transform the basics before playing with the toys.
Myth #3: Profitable companies are the most likely to launch
successful digital transformation projects.
Reality: If things are going well – crudely meaning wealth is being
created for employees and shareholders – the likelihood of
transforming anything meaningful is actually quite low. When
existing market leaders drive disruptive technology, it requires a
commitment from the top-down at every level. It may also require
a significant culture change. At minimum, it will require a clear
operating model to drive the digital and business change.

3Attribution: Stephen J. Andriole is the Thomas G. Labrecque Professor of

Business Technology at Villanova University in Villanova, Pennsylvania
9

..MS2019..
The Digital Wave

Myth #4: We need to disrupt our industry before someone

else does.
Reality: Disruptive transformation seldom begins with market
leaders whose business models have defined their industry
categories for years.
Myth #5: Executives are hungry for digital transformation.
Reality: The number of executives who really want to transform
their companies is relatively small – especially in public companies.
Author’s remarks.
The consequences of ‘do nothing’ can be calamitous. Some boards consider it
wise to elect a Digital Non-Executive Director to challenge complacency.

It is for you to decide if you believe that any of these myths

characterize your organization’s stance. Some businesses operate in
a market that is highly susceptible to change. For them Myth #4
might be highly relevant. Transformation is easier in unregulated
markets. Regulated markets are cushioned by the need for
consultation and legislation. That can inhibit change. In that case,
you have more time to anticipate and consider your move.
Intuitively, the answer for most must surely be: Keep a sharp eye
on what is happening in your external environment (e.g.
competitors) and hedge your bets.

The search for value

The things that qualify as ‘transformation creating value’ in
businesses are often in the areas of cost efficiency and automation.
Or they might lie in being able to offer the customer a more
bespoke product or service. There are other ways that value can be
created in this traditional scenario. All of them essentially come
under the heading of efficiency – doing the same things faster,
cheaper, or better than before.
What is different about value creation in a disrupter model? It is
that it achieves a similar or better outcome to an existing business
model in an entirely different way. Or it delivers an outcome not
previously available. Such disruption is typically a step-change.

10

..MS2019..
 The Digital Wave

 Uber created an app for a customer to hail a taxi online.

Almost instantaneously the customer receives information
about availability and the approximate cost of the journey.
 Netflix delivered new release videos on demand to your
TV. Bye-bye Blockbuster!
 Airbnb allowed a homeowner to rent accommodation
direct to a guest. Some homeowners actually gave up their
jobs to make a living out of renting a room. Wow!
Now that they are successful businesses, they have put others out
of business, or at least made it harder for them to compete.
Thought provoking stuff!
Few businesses choose to disrupt the status quo - until there is no
other choice. Those that do are usually new entrants. Even fewer
established businesses take the risk of disrupting the status quo by
cannibalizing their core business. Other times, what is labelled
disruptive is just fixing something that was badly broken anyway.

Talking business models

The business model is the new challenger ground. The disrupter
starts with a blank sheet of paper. That gives them an advantage
over an established business that has many things to consider
before they disturb their business model. Both have readily
available high-speed scalable and ‘affordable’ Cloud compute
resources. So now business models are under threat from
technology enabled disruption – and you don’t need big bucks to
get started.
It is probably fair to say that, in the case of most new entry
disrupters, their business models have been entirely enabled by
technology. Their definition of success is different. Starting from
nothing and with very low costs they set out to win or steal revenue
and to do it as fast as possible. They innovate fast to stay ahead of
those who are better funded. They often end up being acquired by
one of the incumbents they set out to disrupt.
Those hotheads with an idea working in a garage or a shared
workspace are unencumbered by organizational politics and the
other things that slow you down. They access high speed scalable
Cloud compute resources and clever off-the-shelf code building
blocks. Both are accessible on a pay as you go basis. This is
11

..MS2019..
The Digital Wave

probably the single most important factor in removing barriers to

entry that previously protected the incumbent. Don’t despair; there
are also ways for established businesses to act as challenger. That is
what Geoffrey Moore in his book Escape Velocity 4 calls the
‘asymmetric bet’.
Today new entry disrupters don’t need to read a book about
building a Digital Operating Model. They are ‘born in the Cloud’ 5
and they know no other way. This book is primarily focused on the
Digital Operating Model for the enterprise customer. Their
situation has complexity by reason of their size, existing business
model, and IT legacy. They have the most to gain, and maybe the
most to lose.
Now, many of the business value improvements introduced into
the enterprise so far have not been initiated by people whose core
competency is technology. This is an important observation. These
businesses have not historically been wired in a way that allows
such a thing to happen.
Instead these changes are often inspired by a visionary business
leader. It is they who drive the technologists and the other people
in the business toward an outcome which they instinctively believe
is achievable. They set the challenge. They provide the framework
to facilitate the required change. This is often done by a mandate to
copy or match a disrupter’s new offering. Other times they are just
making a bet. The technologists in the enterprise are then invited to
respond to the challenge of delivering the what and the how.

Supercharge your BOM!

The Business Operating Model (BOM) is difficult to pin down. It’s
what you do and your unique differentiation. It describes how an
organization is run and sustains itself in the process. It is
increasingly reliant on technology. The technology historically sits
with the IT Operating Model (ITOM) to support the organization’s
BOM and aims to meet its needs responsively and cost effectively.

4 Geoffrey Moore is perhaps best known for his work ‘Crossing the Chasm’.
Escape Velocity is relevant reading in support of this book’s themes.
5 Born in the Cloud – a business that from its inception relies entirely on the

Cloud for delivery of its products and / or services and customer services.
12

..MS2019..
 The Digital Wave

However, as more IT transitions to the Cloud then the ITOM has

to evolve.
With the Cloud you can now supercharge your BOM and deliver a
new level of agility with a Digital Operating Model as we describe
in this book. The enterprise IT world is accustomed to a
demand/supply model for innovation – the business demands, and
the technologists figure out how to supply. When it comes to being
truly innovative, this traditional operating model of Business-IT
engagement has major failings. As the business world recognizes
that disruption is probably going to be business as usual – the role
of IT must change.
The IT team must become more entrepreneurial and innovative. It
must develop a more detailed understanding of what differentiates
the business that it is part of. It needs to start making suggestions
about how to be more competitive. Sometimes it should come up
with ideas for: ‘We can save money by doing this better’.
Conversely, the business must get used to being receptive to new
ideas. Some of these may be a little unnerving to live with. Artificial
Intelligence and Robotic Process Automation are good examples.
Those enterprises that have moved into the world of the public
Cloud now have new tools to apply to work. For example, the
speculative processing of large datasets to reveal new insights for
the business to act upon. It is now possible to undertake random
experimentation on a whim that was not financially justifiable ten
or even five years ago. Powerful stuff.
Put simply, it’s reasonable to assert that business executives ‘don’t
know what they don’t know’ about their business and the market in
which it operates. This is not a criticism. It is just an observation. A
trawl of LinkedIn reveals the insights of others on this point.

“ 


The truth is, if you aren't using data appropriately your competitors
probably are.”
“Without big data analytics, companies are blind and deaf, wandering
out into the web like deer on a highway” Geoffrey Moore, author and
consultant.

13

..MS2019..
The Digital Wave

 “It is important for marketing professionals to improve their skills in

data analysis. It takes time and skill to determine which pieces of data
are meaningful. But this process of boiling down data into consumable
chunks is imperative for getting buy-in across an organization.”
 “I hear this guy was pretty smart. At least that's the word on the
street. But seriously, without the data to support a decision, too much
is being left to opinion and chance. Why risk it? Test and Retest.”
 “One of the toughest things to do with analytics is creating a narrative
that is easy to understand for colleagues and executives. Obviously, we
know that visuals are important, but it takes more than that to get
the type of buy-in you want. When deciding which pieces of data are
the most meaningful, start from the end and work backwards. What
am I trying to achieve? What does the audience need to know?”
 “The amount of information available is overwhelming. So much so,
that managers and executives don’t have the time or ability to focus on

”
specific data sets.”

The cloud is awash with innovation and resources for a ‘propeller

head’ (yes, some business teams still think that is what IT are, even
though they are highly skilled, highly paid and high in demand).
These folks are down in the basement ready to cobble together all
the data that the organization has held for all time. Set them to
work on that data – in the Cloud. They will throw it at a ‘pay by the
second’ public Cloud data analytics platform and attempt to
discover something completely new about the business that the
leadership simply did not know.
For example; ‘Did you know that our most profitable customers
are not the ones we think, they are actually those that are …! Or;
‘Our fasting growing client demographic is actually … - perhaps we
should focus our marketing on that small but growing niche
market?’
It is also now possible with the concepts of Agile software
development and advanced coding tools, to test ideas quickly and
at low cost. A failed outcome is acceptable as every outcome can be
a learning opportunity. It is the new way and may feel
uncomfortable at first. The IT team should be saying; ‘Hey, we
could knock that up on our website in a couple of weeks if you
want to. Why don’t we put it out there to see what happens?’

14

..MS2019..
 The Digital Wave

So, the lessons to learn from the business viewpoint are that:
 You DON’T have to be technical or understand digital
jargon in order to be able to demand and achieve the
delivery of value to your business in a digital age.
 You DON’T need deep pockets, simply access the
economics of the Cloud.
 You DO have to create an environment whereby the seeds
of such change are nurtured and encouraged.
 You DO need to set the conditions for your IT
professionals to develop an entrepreneurial mindset.
In a digital age, technologists within an enterprise are key players.
They must be fertile in thinking of ways to support the business to
continuously innovate and work smart.
Work smart in a digital age takes on a new meaning; blending a culture that
promotes ideation with use of technology to test new ideas.

Why should so many of the highly disruptive ideas come from

start-ups? There is no reason. The response of many established
businesses is to create an environment for new ideas in Innovation
Labs. That is the way to go, provided it is run with the discipline of
a start-up! What does that mean? In the cloud! Work Lean. Work
Agile. Readily accepting failure (no blame) and learning. Praise and
reward the breakthrough.

People in the mix

The business that puts its IT in the Cloud can radically change with
the overhead of operating IT substantially removed. This can mean
fewer people are required and/or people with different skills are
required. Then there is the need to consider what key roles are
needed in-house to drive the Digital Operating Model forward.
This can present a dilemma as to who can step up to this role.
Alternatively, do you go outside to find those skills? You can
switch on technology really quickly, but the people in the mix
require more time to come to terms with change and to acquire
new skills and experience. With digital at the core of the Business
Operating Model it is wise to keep a core tech competency in-
house. This is a long game! We talk more about how to structure
your teams in Part 2 of this book.

15

..MS2019..
The Digital Wave

16

..MS2019..
 The Way Ahead

Chapter

3
3The Way Ahead
There are no constraints on the human mind, no walls around the human
spirit, no barriers to our progress except those we ourselves erect
Ronald Reagan (United States President, 1911–2004)

I F only you could press the pause button as you ponder the
seemingly endless assault of articles foretelling unwelcome
news! They tell you your business model is going to be
disrupted. They ask: What is your digital transformation plan
and are you on top of your customer experience? The robots are
coming. . . and. . . and. . . Oh! Then there is the 4th Industrial
Revolution (4IR) characterized by a fusion of technologies that is
blurring the lines between physical, digital, and biological spheres.
(Isn't Wikipedia wonderful!).
There is a lot happening and much head scratching about what this
all means.
We are in a time of boundless innovation and it can send the head
spinning with questions about its relevance. Is it a fad? Will it pass?
Many believed the Cloud was a fad – something that would peter
out. How wrong they were!
Even so we are left with choices. Do nothing is always an option.
Otherwise make a bet, preferably a calculated bet. Every industry
has a predisposition to adopt technology, for example:
 The retail industry has bet on e-commerce. If you are a
retailer today and not online then the future might be bleak.
 The automotive industry has highly automated supply chain
processes and made extensive use of robots.
17

..MS2019..
The Way Ahead

 The financial services industry, for so long rooted in

conservatism, has its own FinTech innovation. They aim to
compete with and disrupt long-standing (and some say,
out-of-date) practices.
 Governments are not usually in the vanguard of technology
adoption. Even so, they have implemented ‘Cloud-first’
policies.
Predicting the future is for the brave - or perhaps the foolish. The
foundations of the future already exist. They are served up as the
4th Industrial Revolution and the technology drivers of that
revolution are illustrated below.

Figure 2 - The 4th industrial revolution

Notice that normal English words are written below the technology
labels in the diagram above. For example, "The eyes" is under IoT:
(Internet of Things). So now we have to think about what that
might mean in a practical world. The quotation at the beginning of
the chapter is especially pertinent.
There is great scope for creative thought to apply technology to
work. In some cases, we now know that human work is likely to be
replaced by a machine. The role of Artificial Intelligence with its
human label, 'The Brains', is an example of technology providing
both opportunity and dilemma. This has caught the attention of
politicians because of the societal impact and ethical
18

..MS2019..
 The Way Ahead

considerations. Microsoft’s evidence to the UK’s House of Lords

Select Committee on Artificial Intelligence 6 is well worth reading.
Business leaders also face challenging times balancing the desire, or
perhaps necessity, to put technology to use in a socially responsible
way. If you think that seems odd, then consider the difficulties that
Google 7 and Facebook 8 have faced from regulators and
politicians. With technology so deeply ingrained in society it has
come under intense scrutiny. With the growing use of technology
there is a new aspect of social responsibility and ethical conduct for
the leadership team to consider.

What lies ahead?

According to Professor Klaus Schwab, founder and executive
chairman of the World Economic Forum and author of The Fourth
Revolution and Shaping The Fourth Revolution (foreword by Satya
Nadella).
This Fourth Industrial Revolution is, however, fundamentally different. It is
characterized by a range of new technologies (see diagram before) that are
fusing the physical, digital and biological worlds, impacting all disciplines,
economies and industries, and even challenging ideas about what it means to
be human.
The resulting shifts and disruptions mean that we live in a time of great
promise and great peril. The world has the potential to connect billions more
people to digital networks, dramatically improve the efficiency of organizations
and even manage assets in ways that can help regenerate the natural
environment, potentially undoing the damage of previous industrial revolutions.

This describes a revolution of great potential magnitude. No doubt

there are consequences we cannot foresee. Does it all seem far
removed from the daily grind of business? Your call.
You may have been an early adopter of Cloud or just getting
started. The evidence is, Cloud is having a transformational impact
on businesses. Cloud looks set to be the future direction of travel.
How would you choose to talk about each of the technologies?
How relevant are they to your future?

6 https://data.parliament.uk/writtenevidence/committeeevidence.svc/evidencedocument/
artificial-intelligence-committee/artificial-intelligence/writtem/69654.html
7 https://europa.eu/rapid/press-release_IP-19-1770_en.htm
8 https://www.ftc.gov/news-events/press-releases/2019/07/ftc-imposes-5-billion-penalty-

sweeping-new-privacy-restrictions
19

..MS2019..
The Way Ahead

It would be a brave business leader to announce how they are

going to deliver on the 4th Industrial Revolution. Maybe we should
leave this to academics for the time being. Is it more realistic to
frame that in terms of, as we describe it, a Digital Operating
Model? Here is a reminder:
The Digital Operating Model is where everyone comes together, Business
teams, Developers, IT Operations. The Digital Operating Model is where
discussions about ‘digital at the core’ of delivering business outcomes get
turned into actionable plans.

Watch this space

The buzzwords ‘Digital Transformation’ and ‘Business
Transformation’ are both popular. Both apply to the way
businesses are reinventing the way they organize their work.
What is Business Transformation? 9
Business transformation involves making fundamental changes in how
business is conducted in order to help cope with shifts in market environment.
(Source: Wikipedia).

Another point of view from the Global Centre for Digital Business
Transformation report: 10
“organizational change is the foundation of digital business transformation”

That’s because changing the nature of an organization means

changing the way people work, challenging their mindsets and the
daily work processes and strategies that they rely upon. While these
present the most difficult problems, they also yield the most
worthwhile rewards, allowing a business to become more efficient,
data-driven and nimble, taking advantage of more business
opportunities.
Turning ideas into action is what matters and this book serves the
recipe. Who is on point for the transformation agenda?

9 https://www.weforum.org/about/the-fourth-industrial-revolution-by-klaus-schwab
10 https://www.imd.org
20

..MS2019..
 The Way Ahead

Executive Line-of-business Built on third

Mandate led platform
• Digital transformation is • Line-of-business (LOB) • The new generation of
at the forefront of leaders such as technologies - big data,
customer conversations marketing, operations, analytics, social and
and is a board-level sales, HR and finance mobile - are necessary to
initiative. Executives leaders are being called enable the achievement
view the path to upon to drive business of a company's digital
capturing this model changes and goals. These
opportunity as becoming technology decisions, technologies along with
technology-centric to and to facilitate digital accelerators like the
deliver new customer innovation toward a Internet of Things (IoT)
experiences. technology-centric and artificial intelligence
business. (AI), will dominate the
unprecedented increase
in future technology
spending.

Figure 3 - Transformation agenda roles

At the front of the book, ‘Who should read this book’ describes the
roles played by all those involved in the transformation agenda.

The idea factory

The new business dynamic favors building teams that are agile.
They need a fail-fast mentality while keeping overheads low. The
Cloud is highly suited to support this way of working. It is equally
applicable to the micro business and to the Goliath. This is both
opportunity and threat, particularly so for the Goliaths. Everyone
will know a business that is a Goliath today but started out with
just a few folks with an idea and a garage. In the same vein, some
Goliaths that appeared permanent, have, er, gone.
New ideas are increasing translated into software and the Cloud has
made that easy. You may have heard the expression ‘Software Is
Eating The World’ 11. It may sound exaggerated until you consider
how dependent we are on software in our own personal lives
compared to ten years ago. Not convinced? How many apps have
you installed on your smartphone? In organizations, software is
now the DNA - it is that important!
The Cloud serves up all manner of tools for transformation. You
serve up the ideas and get down to work. The work is in your
hands and the reference for organization of that work is the Cloud

Marc Andreessen penned his famous “Why Software Is Eating the World” essay in
11

The Wall Street Journal in 2011.

21

..MS2019..
The Way Ahead

Adoption Framework. The Cloud is the factory floor on which

ideas quickly come to life!
In the Cloud Adoption Framework for Azure Innovate section, there is
guidance on how to drive business leaders to become citizen developers as
part of an ideas factory. You can do that without dependencies on pro-
developers until you need the scale, and the business models are all proven:
https://aka.ms/adopt/innovate

Mission Critical?
Before riding off into the sunset; what priority should the
transformation agenda be given? Is it mission critical?
Wikipedia defines 'mission critical' as: any factor of a system (components,
equipment, personnel, process, procedure, software, etc.) that is essential to
business operation or to an organization.

How do you begin to justify the answer?

Product development times are getting shorter. So too are product
life cycles. The business model (the authors are fans of Business
Model Generation 12) must respond to competitive forces 13.
Innovation and the capability to react (get stuff done efficiently)
must be incorporated.
Innovation points to: agile, Cloud, software, developers, talent,
partners
React points to: Digital Operating Model, Cloud Adoption
Framework, Agile, partners
It would be remiss, however, to overlook legacy IT. Legacy (or you
may describe as heritage) IT will continue to support many vital
functions. It is often the system of record holding vital data about
finance, HR, and customers, to name a few. The driver for any
discussion about technology is how it supports the outcome
(profit, customer satisfaction, reducing supply chain costs, etc.).
That drives the discussion to deliver the desired outcome with IT.
Do you upgrade the investment in legacy IT and extend its life or
go to the Cloud? No one size fits all! However, do think about the

12 Business Model Generation at: https://strategyzer.com/books/business-model-

generation
13 Porter's five forces analysis at:

https://en.wikipedia.org/wiki/Porter%27s_five_forces_analysis
22

..MS2019..
 The Way Ahead

platform that will take advantage of the technologies mentioned

earlier.
This book is written for you whether you are either struggling with,
or like the idea that Cloud is going to be the principal actor in
support of the Digital Operating Model. The organizational
capability you need is going to be different from that of the past.
This book is not a textbook for the 'transformation' of a business.
Rather it deals with the realities of a mega trend that involves a
shift to the Cloud to access its raft of innovation.

Reality check
Digital is seemingly peppered in every conversation. When did you
last create a new product or service and not consider what IT was
required to support it? The digital part conversation has gravitated
to Cloud computing and lays down a challenge to the past order of
'own and operate'.
Becoming a Digital Business
“Every company is a software company. You have to start thinking and
operating like a digital company. It’s no longer just about procuring one
solution and deploying one. It’s not about one simple software solution. It’s
really you yourself thinking of your own future as a digital company”
Satya Nadella CEO, Microsoft

It is hard to imagine an organization without IT. But how about

describing your organization as a software company? Is it
comfortable to a talk about your organization as a digital business
with a Digital Operating Model?
 How do you feel about that?
 Is it alien or hard to accept as a reality?
 Are you ready to be fluent in this new language?
 If digital were core to your organization’s existence, what
would need to change?
 What is your capability to develop software?
For many the reality is past investments in IT will survive until it
no longer serves its purpose. There may be justification to ‘lift and
shift’ some past investments into the Cloud if that is the right thing
to do. And sometimes there is no reason for change. The path to a

23

..MS2019..
The Way Ahead

digital business is going to be work in progress for some time to

come.

We have moved away from thinking about data processing

(hardware focused) and business processing (software focused) -
and constraints. We have moved towards thinking about – ways to
quickly respond to the changing needs of the organization - find
the solution. That is the pursuit of agility. Then the conversation is
about making it happen and that is framed in the Digital Operating
Model and the Cloud Adoption Framework.

24

..MS2019..
 Digital Countdown – The Race is on

Chapter

4
4Digital Countdown – The Race
is on
By 2020 every business will have set out on a path to becoming either a digital
predator or digital prey – which will your company evolve into?
Forrester Research

I T is generally accepted as truth that today technology

supremacy creates competitive differentiation. It is also
acknowledged that this is not easy to achieve. Cloud
computing makes technology highly accessible and an
economic alternative to the capital-intensive upfront cost to own
and operate IT. Cloud computing is a now clearly a mega trend. 14
Previously if your organization wanted access to innovative
technology it had to write a big cheque and choices were limited.
Now Cloud is a magnet for innovation and a shop window for
software developers. It presents a showcase of technologies that
are suited to organizations of all sizes. It is low-risk, with free trials
and instantaneously accessible turn on-off compute and storage
available. Cloud is the place to go to search for innovation.
Cloud is creating new household names. Some – you will know
who they are – have risen to fame and fortune. For some, this
mega trend has been a highly destructive force leading to the
demise of organizations – you will also know who they are.

14Mega trends are global, sustained and macro-economic forces of development

that impact business, economy, society, cultures and personal lives thereby
defining our future world and its increasing pace of change.
25

..MS2019..
Digital Countdown – The Race is on

You are a software business

That is not obvious until you think about the dependency on
software and that is partly because it is invisible – unlike digital
disruption. Many are now coming to the realization regarding just
how important internal software development will be. In this
context, software development includes Data Analytics, Business
Intelligence, Robotic Process Automation, as well as other
technologies you build yourselves. These will be key to the
differentiation of businesses in the future.
A change is happening. With software readily accessible in the
Cloud, the question being asked is: Why develop software in-
house? We will answer that shortly. Cloud is serving organizations
with just about everything needed for the must do things in business,
such as running e-mail, accounting, Human Resources, Customer
Relationship Management, and the list goes on.
Yet this ‘go to the Cloud for everything’ approach misses one
major point. The Cloud is accessible to everyone and so creates a
level playing field. That being the case, how do you then create
competitive advantage through differentiation?
This may not be a concern. On the other hand, if software is going
to be a core asset to your business and your Digital Operating
Model then you will want more control of that. The previous
chapter referred to every business being a software business. The
interpretation of that is, software is the source for delivering
productivity, efficiency and opportunities for innovation. With
software at the core of a digital business – and growing in
importance – what decisions do you face?
We are at the point of acceptance that almost all innovation will
come from technology – and specifically from software
development. So, isn’t it now essential for a business to undertake
at least a small amount of software development? Even if it is just
to analyze what your data is telling you about the buying patterns of
your customers, and deduce how to serve your customers
better/faster?
It is not a fallacy to think that whatever business you are in, you are
also a software business as software is Digital Operating Model
DNA.

26

..MS2019..
 Digital Countdown – The Race is on

Buy, Build, or Both?

For a long time, there was a choice between two ways of acquiring
software. Either you bought Commercial Off The Shelf (COTS)
packages or built your own bespoke solution. We long ago
developed in-house enhancements to third-party software. The Buy
and then Customize option is really a special case of the buy solution.
The pros and cons of these alternatives have been much debated.
We won't dwell here on ongoing software lifecycle costs and
managing legacy. Now there is a new option: That is: No (or Low)
Commitment Pay-As-You-Go software licensing models. SaaS
provides a great starting point. Massive amounts of ‘above the IaaS
line’ functionality has recently become available. This is a key
feature of public Cloud platforms. The combination of these two
elements enable what could clumsily be called a ‘Rent and
Configure the Building Blocks’ model. 
Suppose your thinking is to deploy public Cloud as a way to
decommission your datacenter. In which case, don’t overlook the
opportunity to decommission legacy software. There is also the
opportunity to change the focus of future development functions.
Today and in the future in-house coding will be more like gluing
together these off-the-shelf building blocks. 
What is being done here is the orchestration of Cloud services with
the implementation of the Digital Operating Model. To what
purpose? Let's use a specific example.
The COO makes an ask of the IT team: 'I want to analyze a dataset
and I need to know what it can tell me, and fast!’. The IT team take
the dataset and clean it up inside Azure Data Factory. Then they
put it into an Azure Data Lake. From there they build a cube to
present it and visualize it using Azure Analysis Services. They then
sit with the business lead and reveal the information (not data!)
using Power BI.
This has been achieved using four different PaaS / SaaS services
working in harmony and with nearly zero coding. The sum of the
parts is completely customizable and tunable. Some of the language
here is tech-speak. Anyway, now you know, you can impress your
tech team! The point is that the IT team now have available the
tools to rapidly respond to the asks of the business. Like?

27

..MS2019..
Digital Countdown – The Race is on

This has huge implications as Microsoft now does the heavy lifting
for you. New roles are created for architects and their job is to keep
control of the plan for the digital house you are building. There are
new roles too for business-savvy bricklayers who know how to
cement the pieces together in order to extract value from your data
- this is going to be a high in-demand skill!
We have described how the boundaries between IT and the
business are being broken down. The purpose has already been
called out – agility. How to apply this to your organization is for you
to decide.

What is on the shelf?

Answer: Just about everything you can think of, and a whole lot
more you didn’t know existed.
The author’s definitions are below, although there are more
detailed (and technical) explanations at:
https://www.nist.gov/publications/nist-definition-cloud-computing
Software as a Service (SaaS)
SaaS applications revolve around user and application functionality.
They are shrink-wrapped applications, delivered remotely, priced
per user, and billed monthly. They typically provide customization
and configuration options to allow you to align them to the
demands of your organization.
They are though, by definition, pre-packaged. Not everything can
be customized. They do have fixed, predictable costs and very little
(if any) maintenance requirement. Office 365 is a good example.
You bring the users. Microsoft does the rest.
Infrastructure as a Service (IaaS)
IaaS revolves around the server. IaaS has a lot of similarities to a
traditional on-premises, virtualized environment. You can run
servers in much the same way as you do today. The difference is
that the server is in a Cloud provider’s datacenter. The major
difference from what you have today is probably in the level of
automation and scripting made available. You also have, for all
intents and purposes, unlimited capacity to create and delete virtual
servers on the fly. You have complete control over configuring the
operating system, storage, networking, and more. Think: Azure
28

..MS2019..
 Digital Countdown – The Race is on

Infrastructure Services. You bring the Operating System. Microsoft

manages the low-level hardware and virtualization tiers.
Platform as a Service (PaaS)
PaaS revolves around custom-developed applications or application
components. You can build an application by stitching together
several platform services and deploying code on top of them. You
have complete control at the application level, i.e. what the app is
and how it functions. But you have less control of the mechanics
of how that application is deployed and managed. The Cloud
vendor does that for you. This also means you have less to
maintain, support, and worry about. Think of Azure platform
services such as SQL Azure and Azure Kubernetes Service. You
write the code and deploy it as a managed application. Microsoft
does the rest.
Functions as a Service (FaaS)
FaaS is an evolution of PaaS, with even less control, and even less
management responsibility. It provides the ability to execute raw
code as a service. With PaaS you still need to be aware of and pay
for a (web) server, for instance. But you don’t have to deploy it or
manage it. With Functions, you just deploy your code block
(sometimes called a microservice). Microsoft charges you based on the
resource used each time your code executes. Think of Azure
Functions. You bring the raw code. Microsoft does the rest.

29

..MS2019..
Digital Countdown – The Race is on

Figure 4 - Azure Service Map

30

..MS2019..
 Digital Countdown – The Race is on

What does this deliver?

FaaS is still relatively new. The three other services are extensively
used today. In some cases, organizations have gone all in. They
have completely removed their on-premises equipment, replacing it
with a mix of these four capabilities. Those old server rooms have
become meeting rooms! Or a higher profit margin.
Taken together, Cloud services deliver on-demand functionality.
For example, they enable infrastructure to grow on demand and
shrink on demand. You only pay for what you need. Previously
such costs would have been absorbed as a fixed cost to meet the
highest forecast volume of transactions. And then they would have
been on your books for some years, irrespective of how your
macro climate changed.
These capabilities can be exploited as part of an Agile approach to
software development. They are for the creation of value and
competitive differentiation. They are also the tools for the disrupter
to go to work. FaaS moves us firmly into the future world of
‘Cloud application engineering using microservices’.
Now is the time to press the pause button on tech-speak until
Part-2.

Who controls the cost?

There is a widely recognized and obvious downside of allowing
techies (and business people too) to have unlimited access to an
on-demand Pay-As-You-Go computing resource. Business
managers fear unexpected invoices for services, causing havoc for
budgets and cash flow. This is probably right up there among their
top three Cloud fears.
As a result, the emergence and adoption of public Cloud
governance systems is vital. Organizations want to be able to
budget for their projected consumption costs. They want to bill the
time transparently to the incurring division or department. They
want to pre-emptively sign off on limits and breaches. In other
words, they want it to run in the same way as the traditional things
that finance, IT, and the business have done (or wanted to do) for
years.

31

..MS2019..
Digital Countdown – The Race is on

Some organizations opt for a multi-Cloud strategy. Note the

warning in Part 2 on that matter. Having made that choice, they
need tools to help them pre-emptively determine the best execution
venue for their production Cloud workloads. Later we touch on cost
and quality governance in an Agile development world.

The business head conversation

A different focus is needed in the business conversation about
adopting Agile development and using Pay-As-You-Go public
Cloud services. This conversation must center on the view that the
services described above are enablers for new ways to get ahead.
At the same time, it’s vital to keep an eye on mega trends as they
shape the world. The price of ignoring them can be extinction.
Think about the retail businesses that ignored the trends for online
shopping and on-demand movies? Some technology companies
thought the Cloud was fanciful. They said it was nothing more than
a return to the days of the mainframe. How wrong they were!
The business head conversation should not be wrapped up in
technology. It should focus on staying relevant. In his book, Hit
Refresh, Satya Nadella, CEO of Microsoft, calls out four initiatives:15
1. Leverage data to improve the customer experience
2. Support employee productivity with mobile collaboration
tools
3. Optimize, simplify, and automate business processes
4. Transform products, services, and business models
Note how step 4 cycles back into step 1 in an iterative manner.
Turning those initiatives into practical solutions is the next step.
We now bring into focus some mega trends that are shaping and, in
some cases, worrying society. 16

15 From Hit Refresh (ISBN 978-0-00-824765-2, page 126)

16 The US Government set out its position in a paper (May 2018) at
https://www.whitehouse.gov/briefings-statements/artificial-intelligence-american-people/
32

..MS2019..
 Digital Countdown – The Race is on

Artificial Intelligence (AI)

It is early days for AI, but there are lots of examples that you may
have already been exposed to. For example, have you visited a
website and seen a pop up with a chatbot? That is a practical, basic
use of AI.

Big Data (BD)

It is not often the case that we have too little data. More often
there is too much to make sense of in order to provide the insights
we seek. Somewhere buried in the data are nuggets of insight. If
only they could be exposed. You could put people to work on what
is frankly a thankless task. A better way is to use a computer to run
n scenarios in parallel in a matter of minutes. That is big data at
work.

Robotics Process Automation (RPA)

Factories around the world are advanced in their use of robotics.
Now RPA is coming to the office where it can automate routine
and repetitive human/computer interaction tasks. Insurance firms
are beginning to staff the e-mail response agents in their call
centers with robot-like software. This can read, process, and
respond to common e-mail queries without human involvement.
That is RPA at work.
Cloud is the enabler, and so building a core competency in Cloud is
fundamental. The Cloud Adoption Framework is there to guide
you. Be ready and prepared to use these capabilities when they
become relevant to your organization. You are reading this book;
you are on your way!
Now you may or may not be ready to put these capabilities to work
immediately. Check out the noise with your familiar networks to
gauge what is going on. You might not want to be “first off the
block”. On the other hand, last is never a good place to be.

33

..MS2019..
Digital Countdown – The Race is on

Don’t get caught out

The amount of data being created daily is truly astonishing. The
Cloud has ample capacity to deal with this serving up massive on-
demand computing power to store, index, and process. The pursuit
of an improved business will necessitate finding better ways to
process and analyze data. In short, finding ways to extract
information from the ever-expanding data sprawl. There is no
shortage of ideas for how that can be done. It is a big task
involving more than just technical challenges. A more recent
challenge is coming to terms with what you are allowed to do with
that data.
As you investigate the possibilities to put the technologies referred
to above to use, it is vital to understand regulation governing the
use of personal data. In Europe the General Data Protection
Regulation (GDPR) came into force on May, 25 2018 and
substantially upgraded the rights of EU citizens (c. 513M) to
control how their Personal Identifiable Information (PII) is used
and curbed the freedoms of businesses to do as they like when
processing PII.
There is a growing concern among consumers about how the
personal data they share with businesses is used. Europe leads the
world with its General Data Protection Regulation. India is seeking
adequacy status (permitting the lawful transfer of data between EU
member states and India) under the European Union’s GDPR.
Andorra, Argentina, Canada (commercial organizations), Faroe
Islands, Guernsey, Israel, Isle of Man, Japan, Jersey, New Zealand,
Switzerland, Uruguay and the United States of America (limited to
the Privacy Shield framework) all have adequacy status (as of the
date of publication of this book). The California Consumer Privacy
Act is expected to go into effect Jan. 1, 2020. More will surely
follow.
You will do well to tread carefully with the adoption of technology,
such as Artificial Intelligence, for the processing of PII. Data can
be stored, processed and transmitted anywhere in the Cloud. Your
Digital Operating Model is not geographically confined, unless you
decide that. So, always check compliance requirements with
prevailing law/regulation. And do keep in mind that compliance is
not globally harmonized. Don’t stall out of fear, rather recognize
the importance of due diligence.
34

..MS2019..
 Digital Countdown – The Race is on

On the one hand technology is becoming ever more capable of

processing data and extracting the insights you seek. On the other
hand, you are more tied as to how you are allowed to process data.
If this is of importance to you then you are recommended to read
the guidance 17 published by the UK’s Information Commissioners
Office (ICO) relating to AI, BD and RPA. That guidance topped
the votes in the People’s Choice Award at the 39th International
Conference of Data Protection and Privacy Commissioners
(ICDPPC).

Undercurrents
Technology is so ingrained and important to society that it is on
the agenda of governments. This is a recent phenomenon. There is
growing interest from policymakers in regulating the technology
sector. At the same time, they want to avoid stifling innovation.
Should this put a stop to your plans? No.
The very institutions that policy makers work for, or are elected to,
have similar challenges to businesses but on a much greater scale.
They need technology to advance their work.
Even so it is wise to recognize that the techno-political debate is
becoming louder. If you work in the media industry you will know
that only too well. Keep a watching brief and talk to Microsoft.
Because of their size and global influence, they are highly active in
the techno-political debate.

You are a software business?

Driving the message of this chapter home:
 Do you treat IT as a priority for investment in the race to
get ahead?
 Do you see value in building the capability to be a software
business?
(Software as the source for delivering productivity,
efficiency and opportunities for innovation)
 Do you want to enable your teams to react quickly – to be
agile?
17https://ico.org.uk/media/for-organisations/documents/2013559/big-data-ai-ml-and-
data-protection.pdf
35

..MS2019..
Digital Countdown – The Race is on

 Do you have in place strong governance concerning the

processing of data?
 Do you know which partners you want to take on your
journey?
We hope this chapter has elevated your attention to these things.
The extent to which software development will differentiate your organization
from your competitors will vary greatly across industries. It will depend on your
value proposition and how software supports that. If you are in one of the
industries most likely to be disrupted, and you haven’t invested in Agile yet,
you had better read on.
The gun starting the race has already been fired!

36

..MS2019..
 Agile – small word quick work

Chapter

5
5Agile – small word quick work
Perfection is not attainable, but if we chase perfection, we can catch excellence
Vince Lombardi (American Football Player, 1913 – 1970)

D EPENDING who you ask what agile is, they will tell you
it is about moving quickly, or, they’ll probably tell you that
Agile (with a capital A) is to do with software
development and that it’s particularly helpful for software
development projects with an innovation agenda. Both are correct.
Here we refer to Agile. For the leadership team the thing to know
is; if you want to be agile then Agile makes that happen.
The key difference between Agile and the standard coding
methodologies for projects that preceded it is that Agile has as
concepts:
 short term bursts (sprints) of coding
 regular face to face meetings (daily stand-ups)
 small teams that have a range of specializations
 test, fail quickly, avoid wasted effort and speedily change
direction
 coordination of small coding activities in parallel (rather
than big ones in a serial sequence – hence the dominant
predecessor’s name “Waterfall”).
Agile has for the most part become the de facto mechanism for
software development in recent years. You can do Agile coding
without DevOps and/or Cloud. You can use DevOps and/or
Cloud without Agile. They just happen to have evolved in broadly

37

..MS2019..
Agile – small word quick work

similar timeframes and they co-exist in an organization or a project

very well.
It is not our goal to explain Agile development to you. What’s
interesting about Agile within the context of this leadership Part of
the book is to explore its impact on the organization as a part of
the IT Operating Model/Business Operating Model/Digital
Operating Model discussion, and to position how Agile is about so
much more than just software development. Agile and its bigger
brother, Scaled Agile Framework® (SAFe®) for Lean Enterprises,
are the glue that blends these three operating models together.
Why did Agile evolve into what it is today in the application
development space?
Agile aligns today’s software developers with the needs of the
business that is continuously in search of innovation. It is fast.
Much faster than methods used previously, and it is that speed that
supports the business to respond to change. Whether that change is
planned, or in response to external events.
We live today in a business world of ‘accelerating change at an
accelerating rate’. In this new world, horizon scanning and taking a
long-term view of what innovation will keep the business
competitive, is simply not viable.
The ability to react to the availability of new innovation and do so quickly is as
important to being competitive as the innovation itself.

Read that again as it is fundamental to Agile’s raison d’être.

Innovation in the Cloud is continuous and available to all-comers.
In times gone by, those with the resources (deep pockets, skills)
would outgun others. That is no longer the case. The competitive
landscape has been levelled and actually tipped in favor of those
that do not have the burden of legacy systems to maintain. Do you
see the problem for some?
The central tenant of Agile is the backlog. Think of it as a long
wish list. Every couple of weeks you pick new things from your
wish list and deliver them. Because you only ever pick things off
your wish list every couple of weeks, you can change tack much
more quickly to respond to changes in priorities and external
factors. The backlog is the central reason why Agile is so … agile.

38

..MS2019..
 Agile – small word quick work

Agile budgeting and forecasting

One of the biggest challenges with Agile is how you go about
implementing it. Delivering small, discrete projects with Agile is
relatively straightforward as it doesn’t profoundly impact how you
budget for these initiatives or require you to make wholesale
changes to your processes. To truly embrace Agile, you need to
make a few more major changes to the way you think about
projects, budgets, and deliverables.
Traditionally, both the business and IT tend to think about IT
projects through the lens of applications and projects. We need a
new application, or we need to change an existing application, let’s
create a project and align resourcing and budgeting around this
project. This is the wrong approach.
Instead, organizations that are truly embracing an Agile approach
are doing things a bit differently. Instead of applications and
projects, start to think about the notion of a product. This may be
an internal or external product. Drop the functional specifications
and handoffs between business and IT.
Instead, assemble a cross-functional team, comprising business
experts, developers and other members of the organization in one
cohesive, stable unit. Cross the business and IT chasm.

A large medical-device manufacturer significantly shortened it’s time to market

by refining its organizational structure. Under its traditional structure, there
could be as many as 20 handoffs when a business unit shared its
specifications and requirements with the technology organization for a new
piece of software or an additional feature in existing software. Because of the
interdependencies among its products, leadership knew it wouldn’t be enough
to deploy agile within one business unit or within certain product-management
teams in the technology organization. In 2015, the company tweaked its
product-ownership model so that software requirements were directly
transmitted from dedicated product owners in the business units to the agile
teams, rather than passing through multiple parties. With the change, the
company was able to reduce the amount of time it took to release products in
the market. The structural changes also facilitated the rise of several
communities of practice. These role-based or topic-based groups (sometimes
called guilds) are critical in agile-at-scale environments. 18

18https://www.mckinsey.com/business-functions/digital-mckinsey/our-insights/an-operating-
model-for-company-wide-agile-development
39

..MS2019..
Agile – small word quick work

Getting the team structure right and changing the culture is but one
aspect of the transformation you must go through. One of the
most difficult changes to navigate is that of budgeting and financial
planning. Historically, IT budgets were set every year and there
were often painful trade-offs between different projects and
initiatives the businesses were driving IT to deliver. Instead, align
these budgets to these product domains and empower product
owners to direct this funding throughout the budgeting cycle,
having regards to changing business priorities.
The Disciplined Agile Consortium 19 has lots of great content to
help with formalizing this budgeting process:

Figure 5 - Agile funding models

An interesting concept which is becoming more prevalent is the

concept of venture-funding product development internally. In the
same way a team of entrepreneurs might approach external parties
to fund new products and services, now internal teams can vie for
funding from internal investors. Initial funding might cover an
MVP and then, directly linked to the success of the product with
employees or customers, additional funding can be made available.
This interesting approach laser-focusses budget in line with success
and therefore significantly reduces wasted expenditure and cuts
development running out of control.

19 http://www.disciplinedagiledelivery.com/secure-funding/
40

..MS2019..
 Agile – small word quick work

Whatever approach you take to Agile budgeting and planning,

ensure you major on flexibility and adaptability, whilst maintaining
close visibility and overall control of your all-up spend.

Agile for the Enterprise

Agile is about so much more than just software development
though. Agile has started to permeate just about every part of the
business and has become a useful tool to manage almost any
project you might want to run. What’s interesting is how Agile can
become the mechanism you use to run your most important
project of all - your Digital Operating Model. This is where SAFe®
comes in.
So, what is SAFe®? Several years ago, Dean Leffingwell wrote a
book called Agile Software Requirements 20. It became a best-seller that
gave the world its first glimpse of what would eventually be known
as SAFe®. It described a framework – called the “Agile Enterprise
Big Picture” – that visualized how to apply Lean and Agile
practices and principles to the Team, Program, and Portfolio
Levels.
The concepts behind the Big Picture drew from the knowledge
pools of Lean, Kanban, Scrum, and Extreme Programming (XP),
as well as Don Reinertsen’s The Principles of Product Development Flow.
The distillation of these concepts into a single framework was
applied in places like IBM, Discount Tire, John Deere, and Nokia
with inspiring results.
Today, SAFe® is in its fourth iteration and has been adopted by 70
percent of the Fortune 100.
In its most simplistic form, SAFe® provides a framework for
managing multiple business-focused streams of work within an
enterprise. A kind of backlog of backlogs. It defines these within
the context of portfolios of epics, or major initiatives. An epic
might be a new product that you want to launch. It might be a
major technology refresh program you want to undertake. It might
be a new market you want to enter. It can be anything.

https://www.amazon.com/Agile-Software-Requirements-Enterprise-
20

Development/dp/0321635841
41

..MS2019..
Agile – small word quick work

SAFe® provides a mechanism to manage and then prioritize these

major initiatives and provide a bridge into both the IT and business
teams that might enable this epic to be delivered.
Underneath the portfolio (epic) level is the program level. Epics are
just that, epic! Each epic is broken down into a series of features
and enablers which are grouped into a program increment, which is
a time-bounded period where you deliver some form of business
value. This value is managed by an agile release train. Within each
program increment, you have a number of scrum teams who
actually deliver the work. Over, you’ll see a diagram showing the
high-level view of this process.
The thing that is most interesting about SAFe® is its applicability
across such a broad range of initiatives you might be running as an
organization. The fact that by using SAFe®, Agile can be applied
to far more than just software development. Agile budgeting and
planning can now be used to represent and manage the entire
innovation pipeline within your organization.
As an IT organization, your responsibility becomes standing up,
feeding and watering a given number of downstream agile sprint
teams. These multi-disciplinary teams will span different
technologies. They will have different skillsets. They will do
different things. But they will be there, willing and able, to work
through the constant stream of requirements delivered up and
prioritized by the business. No more functional requirements. No
more detailed business cases. No more friction between the
business and IT. Peace, and hopefully, harmony.
This brief description doesn’t begin to do justice to the intricacies,
or power, of SAFe®. When you fully understand it and see it in
action, you will appreciate its power and how useful it can be. For
more information on SAFe®, please see www.scaledagile.com.
What we seek to do here is simply to signpost you towards SAFe®
and to start to think about how Agile, applied at scale, can be that
bridge between the business and IT. A bridge between your
Business Operating Model and your IT Operating Model.
Here is the authors’ forecast: The businesses that crack the
implementation of Agile development and the power of public
Cloud will be the businesses most likely to succeed in the next
decade.
42

..MS2019..
 Agile – small word quick work

So, the most important objective for the implementation of a

Digital Operating Model is to create the way for real time
interactive collaboration of the business team with the technology
team. It’s called Agile. It’s that simple.

It all starts with an idea …

We spoke earlier about how, as a company, you need to “disrupt
yourself” to maintain relevance in the marketplace, but how can
this be accomplished? How can we compete with the fresh ideas of
the start-ups?
Consider your advantage: The people in your company and your
existing customer base that can be leveraged. Your employees and
customers have a wealth of experience and knowledge that can be
called upon. Within these people are gold nuggets of ideas waiting
to be implemented! We just need to consider how we can
effectively mine for them.
Therefore, make sure you allow a forum for these ideas to be
captured. This can be achieved through probing surveys (internal
or external) and an ideas portal for example. Such an ideas portal
could allow for voting and alike to measure demand.
Your organization should be able to take onboard these ideas and
pivot accordingly. But how do we sort the wheat from the chaff?
We need to validate there is of course demand as well as value to
the business.
Assessing the value of ideas allows the organization to prioritize the
good ideas for further analysis and possible implementation. Such
value to the business can be ascertained through a number of
different prioritization frameworks such as MoSCoW (Must Have,
Should Have, Could Have, Won’t Have), RICE (Reach, Impact,
Confidence, and Effort), WSJF (Weighted Shortest Job First) or
even just HiPPO (Highest Paid Persons Opinion)!
We then transition these ideas to our backlog in the form of epics,
features, user stories or even just improvements or bugs. With
Agile budgeting in line with the SAFe® framework, we now also
have a value stream that these ideas can be funded by alongside the
previously identified backlog items.

43

..MS2019..
Agile – small word quick work

Figure 6 - SAFe® Portfolio Configuration © scaledagileframework.com

44

..MS2019..
 The Business Questions

Chapter

6
6The Business Questions
Take the attitude of a student, never be too big to ask questions, never know
too much to learn something new.
Og Mandino (American Author 1923 – 1996)

I N recent times the leadership team have been subject to

increasing scrutiny of their understanding and oversight of the
technology that is a foundation of the business. That does not
mean to say they have to explain how the organization has
embraced DevOps. It does mean they need to be ready to talk
about and explain the commitment to build an organizational
capability to put technology to work in this digital age – and the
core of that conversation is explained in Part 1.
We have divided these questions into the same structure as the
chapters within the Part 1 above. If you need some context around
the questions, flick back to the chapter to which they refer to give
yourself a quick reminder on the rationale of the question, or to
help you best answer it.
There are no right or wrong answers. As this book suggests,
consult others and seek advice from your vendors or from a
business partner. This is important work.
1. Chapter 2 – The Digital Wave
2. Chapter 3 – The Way Ahead
3. Chapter 4 – The Race is On
4. Chapter 5 – Agile - small word, quick work

45

..MS2019..
The Business Questions > The Digital Wave

6.1 The Digital Wave

Have you encountered an organization lately that only deals with
others by phone and post? No. We are in a digital age and it is
expected of every organization that they understand what that
means. We talk about that as a digital wave and it is unstoppable.
The words disruption and transformation are so topical now. What
is your conversation about?
In this chapter above, we discussed the environment that is driving
the agenda of organizations as they consider:
 The Business Model and its relevance in a digital age
 Disruption – what disruption?
 Creating Value – the output of transformation
 The Business Operating Model (BOM)
 The extraction of value from data
We will now give you some questions to ask yourself and others
about these topics.

46

..MS2019..
 The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.1 Has your If you are operating in a sector that is
industry already used to disruptive change (like
experienced Retail or Media) then you are already
disruption from attuned to understanding the impact of
new entry change. If your sector has long-standing
technology-based tradition and/or is regulated (e.g. Finance,
alternatives? Accounting and Legal) then there are good
reasons why it hasn’t had a major impact
so far. That does not mean to say major
changes are not coming. For example,
Retail Banking is now being disrupted by
challenger banks. Developments such as
blockchain, robotic process automation,
machine learning and artificial intelligence
are forecast to turn these services sectors
on their heads. Is that in 1 year, 5 years or
10 years? That is the bet.
⎕ 6.1.2 Is the The availability of on-demand “pay as you
Board on-board go” computing services means that the
with Cloud? business teams’ ideas can be quickly tested
and deployed into production. The
application of Agile, Cloud and DevOps
are fast becoming established structural
pillars in organizations that are innovative
in their use of technology. The speed with
which your organization can test and
deploy new ideas is the goal of achieving
agility. The Board may not understand the
mechanics of achieving agility yet must
understand its importance. Is that the case?

47

..MS2019..
The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.3 How do Before you start to worry if your
you characterize competitors are about to steal your lunch,
the current state ask what you know about your (and their)
of adoption of state of readiness with Cloud computing
Cloud by your and the associated advanced
competitors? For implementation toolsets?
that matter how Microsoft 21 suggest three states of
do you think readiness:
they characterize
you? Foundational: Organizations with little to
no Cloud experience are still envisioning
Cloud and require end-to-end guidance for
Cloud adoption.
Intermediate: Organizations with
foundational Cloud experience. Have an
understanding of Cloud technologies and
are either evaluating or have migrated non-
priority workloads to the Cloud.
Advanced: Organizations with advanced
Cloud experience are in the process of
Cloud adoption and want to optimize
workloads on the Cloud.
When did you last check the available
sources of information that show the
uptake of Cloud computing in your sector?
Are you a leader or follower? Should you
be investing in being a leader?
Note: Cloud readiness and Cloud maturity
(see Part 2, Chapter 8) are different.

21 Extract from Adopting the Cloud Operating Model

48

..MS2019..
 The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.4 How Take a guess. Ask others and compare.
dependent is the What does that tell you?
delivery of your The business model has many moving
Business Model parts and a supply chain and value chain. It
on IT? likely relies heavily on fast and efficient
collaboration and that comes with
complexity. If your business process is
heavily dependent on IT for ordering,
inventory management, delivery and the
like, then you are more likely to be able to
benefit from (or suffer from) IT (Cloud)
related disruption.
Are you actively exploring how your
customers want to be served and delivering
the optimum customer experience, because
if you are not then you may lose out to
others that are attentive to the search for
smarter (including cheaper) ways to do
that.
⎕ 6.1.5 When If your organizational structure does not
did you last support a bidirectional exchange of ideas
review how the with the IT team (and suppliers and
partnership independent advisors), then you might be
works between missing out on a raft of innovation
the business and opportunities. Many of the new functions
the IT team or offered by public Cloud computing
your IT partners? platforms are complex and understanding
the potential they have within your
organization can only be done as an open
collaborative on-going discussion. We have
an idea and want to do this? Why? It makes
it easier for us to serve our customers.
What do you want to change about how
you serve them today? If the answer is
technology enabled, then the IT
team/partner needs to be right in the mix
to advise what is available.

49

..MS2019..
The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.6 When IT is the enabler and deliverer of an
did you last increasingly digital world. Cloud computing
review your makes everything digital more accessible to
competitive organizations of all sizes. What that means
position? is your competitors who may not have the
resources you have, because they do not
have the capital, can now use Cloud to
compete on a level playing field. That is a
game changer. Are you in this new game?
Are you seeing more new entrants in your
sector? Are they having an impact?
⎕ 6.1.7 Do you IT now needs to be tightly linked to
see (and treat) IT everything from strategy to delivery, and
as both a threat Board level representation is now
and a potential mandatory. If your organization still treats
competitive IT as a supply and demand relationship,
advantage? then you may be missing out on the know-
how of your IT team. It is most likely that
disruption will come at you from a
technology-based platform – are you
protecting that flank?
⎕ 6.1.8 Have you If you take heed of where disruption in
considered how many sectors has occurred, it is in these
you might two areas. 1. how clients search and shop
optimize your for a product or service. 2. how
customers’ ‘shop’ conveniently (and sometimes how quickly)
and that is delivered. Delivering a value
‘convenience’ proposition based on time and
experiences convenience differentiation can justify a
using price differential. Technology has served
technology? the pursuit of differentiated value
propositions and will continue to be the
primary vehicle for doing so. Do not
overlook equipping employees with
technology where they support the delivery
of your value proposition.

50

..MS2019..
 The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.9 Do you If you believe innovative ideas can come
have a culture to from anyone then the culture needs to be
encourage able to elevate those ideas to the
innovation? appropriate discussion forums. Some
believe innovation should be centered
within a Lab format with a budget and
start-up culture. You need to enable and
actively promote the propagation of ideas,
trial and error, and reward the effort (rather
than the result) no matter which model you
adopt. Don’t expect overnight results – this
process takes time particularly in large
organizations. The result may in the end be
the one great idea that supercharges the
organization.
⎕ 6.1.10 Do you Organizations collect an ever-expanding
collect data, or universe of data – with an associated cost
do you distil it to retain and protect. The real value of data
into much its is revealed by the science of data analytics
more valuable (aka Business Intelligence). Data is data
cousin, and what organizations strive for is value
information? from that data and extraction of that value
is a function of people and technology.
There is no shortage of technology. A
more fundamental question is: are you
training people with the skills required to
analyze information?

51

..MS2019..
The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.11 Does If not, then there is work to do. You can’t
your technology expect the IT team to contribute to the
team understand innovation agenda if they don’t ‘get’ the
what your basics of the why, what and how of the
business does, organization’s existence. The history of IT
and what makes has pigeonholed technologists (they just do
it different to its tech, don’t they?). Not anymore, they are
competitors? key to unlocking the potential of
technology for transformation of the
‘always done it this way’ and delivering
‘new ways’ to innovate (overused word but
forget it at your peril).
⎕ 6.1.12 Are you Taking a bet to disrupt a business model
set up to think (to be a challenger) is a major decision
and act like a point for most organizations. If you don’t
challenger or prepare for it, history informs someone
more steady as else will. If you want to be a challenger
she goes? what do your answers to the questions
before say about your propensity to be a
challenger?
⎕ 6.1.13 Which of What does that tell you about your
the myths organization’s current stance? Have you
exposed in discussed this with others in the
Chapter 2 did organization and is there a consensus? If
you most readily there is then you may be susceptible to
agree with? ‘group-think’ so, consider inviting
independent external validation of your
stance.
⎕ 6.1.14 How This is one test of your organization’s
would you agility, the capability to react to an event
characterize your with the analysis of data to reveal insights
capability to to support decision-making. Can you think
process data into of an example where decisions were made
information? on a hunch when you really wanted more
information? In a world where governance
is becoming ever more important, what
capability needs to exist to support
governance with high quality information?
52

..MS2019..
 The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.15 When The capability of your IT team to support
did you last the organization’s digital journey is make or
perform a skills break. What skills gaps does this book
audit of your IT expose?
team?
⎕ 6.1.16 Do you Big changes are occurring with the
have a availability of Cloud services and your team
committed need to be ready and equipped to face off
budget for the the challengers to your business model as
Continuing many of them are born in the Cloud and
Professional Cloud savvy. The rate of innovation
Development necessitates on-going CPD and of course
(CPD) of your there are partners as a source of
IT team? knowledge.
⎕ 6.1.17 Is your In the past it may not have mattered.
IT team Today with IT at the core of the
competent to organization’s value proposition it is vital.
present the Those that set out with disruptive ideas
organization’s understand how technology underpins the
value value proposition and that is becoming a
proposition? key attribute of an organization equipping
itself in this digital age.
⎕ 6.1.18 Do you You may ask; why do we need one? If you
have an accept that your current business model
Innovation Lab? may be under threat of disruption, then
where is the source of new ideas that will
spark your reinvention? This potential for
disruption is perpetual so how is that
addressed?
⎕ 6.1.19 What A skills audit reveals your current skills
skills are needed base. Are they skewed to support existing
for your core IT assets? What are the skills you consider
‘tech’ to be core ‘tech’ competencies for the next
competency? five years? This book will help you answer
this question.

53

..MS2019..
The Business Questions > The Digital Wave

 Question Why this matters

⎕ 6.1.20 What is The own and operate model for IT is a
the comfortable place albeit with an overhead.
organization’s There is a gravitation to the Cloud
appetite to use (evidence its meteoric growth) and that can
Cloud services? challenge that comfortable place. Are you
being left behind? If the future is a growing
reliance on Cloud services what are the
implications for you?

54

..MS2019..
 The Business Questions > The Way Ahead

6.2 The Way Ahead

What is the capacity of your organization to absorb change?
What is driving that change?
It seems that technology is at the core of change and the rate of
innovation is hard to keep up with. Still you don’t want to get left
behind so you need a plan and that is bigger than the selection of
technology.
In this chapter above, we discussed the drivers of change and some
of the big implications that weigh on the leadership team as they
consider their responsibility to stakeholders.
Some questions follow to spark the conversation to shape the way
ahead.

55

..MS2019..
The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.1 How Is this a live conversation or background
does your noise? Who is taking the lead; the business
organization team, IT team, or have you (as this book
think and talk suggests) joined up? For some, digital
about digital transformation has become a “catch-all”
transformation? budget winner for all of the pre-existing
change that was already occurring. Others
have decided on implementing one
initiative, e.g. let’s create a mobile App for
our clients. In reality, digital transformation
should be on-going, focused on reinventing
parts of the business model and the way
they are delivered - incremental and agile
(vs. monolithic and large scale).
⎕ 6.2.2 Does the Does the Board recognize the significance
IT function have of the Digital Operating Model? If most
a seat at the top organizations are becoming (in part at
table? least) highly specialized ‘digital’
organizations, then surely it makes sense to
have digital experience represented at
board level. Go one better and appoint
Non-Executive Directors with digital
experience to challenge the status quo.
⎕ 6.2.3 Are you A digital business uses its data as an asset
looking for clues and mines it to reveal hidden nuggets of
in your data to information. There are Business
inform the Intelligence tools now to automate
business about processing of data and visualize as
what actions to information for human consumption. Yet
take? it still needs to be eyeballed by someone to
interpret the data generated. Machine and
human combined. It is fast becoming
mainstream for an organization to develop
this competency. Are you?

56

..MS2019..
 The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.4 Do you Technology is pervasive in our lives and
access the for some it is a hobby or even an
experience of obsession. Younger generations change
your staff? their preferred mode of communication
with increasing frequency, and they seem
to intuitively know which application to
use (this week) for which purpose. They
suppress their inner pride and pass this
knowledge (with much rolling of eyes) to
their older friends and relatives. This latent
skill set of young people in your employee
base is valuable to the organization. Do
you provide a way for your people to share
ideas with an incentive to do so? And do
you listen when they do? Perhaps the new
young hire in dispatch has some great ideas
– will they remain just ideas or maybe
deliver a quantum improvement in the
work of the dispatch team?
⎕ 6.2.5 Go big The word transformation conjures up big
or small? change but that is not the only way. One
approach is to begin a process of micro-
segmentation. Look at the Line of Business
(LoB) application suite, what can you
reimagine and make Cloud native on a ‘as
you go basis’. Enhancing a LoB application
with new functionality? Is there a Cloud
service as an alternative? Another is to look
outside the organization and to learn what
others are doing and consider if that is an
option. To quote Deloitte 22 from their
report Courage under fire: Embracing
disruption; ‘digital should be defined based
on business needs’.

22 https://www2.deloitte.com/global/en/pages/risk/articles/directors-alert-courage-under-
fire.html
57

..MS2019..
The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.6 Is IT If the IT function reports into the CFO,
viewed a Cost then a cost centered IT mindset is most
Centre or a likely. Splitting off a chunk of IT (and its
Profit Centre? budget) and aligning it with the line of
business management functions will likely
yield results that would not otherwise be
available. The business needs to be able to
engage directly with IT. In turn, IT need to
be able to understand the needs of and
feedback directly to the business leads.
Breaking individuals out of IT and
spreading them throughout the business
teams is not new. Some like the more
radical approach of devolving some IT
functions completely to the business and
deploying a well-coordinated matrix
management approach to deliver on agility.
⎕ 6.2.7 Is change Who likes change? We’ve been here before
feared? when Cloud was the hot topic of
conversation for the IT team and thoughts
turned to; “So, what does this mean for my
future?” Putting the IT team alongside the
business team breaks down this fear as the
IT team is integral to the success of the
business team. Maybe for once they also
share in the rewards? No conflict of
interest arises and what needs to get done,
gets done.

58

..MS2019..
 The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.8 How do ‘I hate unsolicited spam emails!’ is a
you set the stage common complaint from workers at all
for the future levels, and some are just downright
you are inappropriate and unwelcome. However,
envisioning? this never-ending stream of approaches
and offers is still one of the best methods
of easy to access competitive intelligence,
new ideas and market data. Attitude to
spam and making time to attend
educational events is an interesting cultural
litmus test. Most employees are simply too
busy to conduct untargeted, open-ended
outbound market research. Should you
provide filtered “commuting time” reading
from sources that are curated under the
direction of the leadership team?
⎕ 6.2.9 Is digital The financial crisis of 2007-8 revealed no
on the agenda of organization is too big to fail. More
every board recently the failure of organizations is often
meeting. linked to the fact their business model did
not evolve with changes in the external
environment in which they operated. In an
increasingly technology driven world the
Board is dealing with a new phenomenon.
Taking ownership of the digital strategy is
becoming a new benchmark of the board’s
competency. Too busy? Then consider
creating a committee tasked to advise the
board on all matters related to the use of
digital.

59

..MS2019..
The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.10 Who There is no right answer to this question.
owns the Digital The Digital Operating Model upon its
Operating implementation becomes an organizational
Model? practice – “It’s the new way we do things
around here”. Supporting it with a Cloud
Adoption Framework will help to
coordinate and organize the organization’s
resources. Implementation - look at how
other practices were embedded in the
organization to see what worked best.
⎕ 6.2.11 What In this book we advocate keeping a core
capacity and ‘tech’ team that is interfaced to the
competencies do business teams and has competency to
you think you translate the needs of the business and
need to develop deliver a solution. Decisions will need to be
software? made as to what software is developed in-
house and what is served up in the Cloud.
How many people and what competencies
are needed are then framed under the
Digital Operating Model and the Cloud
Adoption Framework. When all is said and
done, some central coordination is
necessary for reporting on cost control and
governance.

60

..MS2019..
 The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.12 Who has This role is understanding the art of the
the responsibility possible and fits with the experimentation
to stay abreast of that is so easy to enable in the Cloud. Ideas
what public can flow from the business teams to IT or
Cloud services vice-versa and someone has to be a
can do to translator and up to-date with what is
enhance your available from public Cloud services.
competitive Public Cloud vendors are releasing new
edge? ‘above the IaaS line’ functionality almost
every day. These features and functions
might just present exactly the functionality
you are looking for, so you don’t have to
build it. Keeping abreast of these
developments is important to driving tech-
led innovation. Who does this? Psst, it’s a
full-time job.
⎕ 6.2.13 As the There is always competition for
balance of expenditure and a hurdle metric for
expenditure on deciding which projects get the go ahead.
Cloud services With legacy systems you had to justify
versus legacy capital expenditure and the recurring cost
systems changes of support and maintenance over the
what is the lifetime of the project. With public Cloud it
impact on the is on/off and you only pay for what you
business case? use. However, legacy code still carries
costs. As a quip: If it’s not paying off, then
turn it off!

61

..MS2019..
The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.14 Have you Cloud First is interpreted as ‘the solution to
evaluated if a a requirement must first be met by Cloud,
‘Cloud first’ and only if that is not viable then consider
strategy right for an alternative’. SaaS, PaaS, IaaS in that
you? order. FaaS glues it all together. This
principle is applied and tested against what
is most important to the organization, for
example: Scalability, Resilience, Agility,
Cost, Innovation and Compliance to name
a few. Cloud is not the right answer for
some workloads. Large enterprises support
an IT legacy and the decision as to whether
or not you should ‘lift and shift’ this legacy
into the Cloud is complex and certainly
non-trivial.
⎕ 6.2.15 Do you It is not always the case that everyone will
have (or are you) be on-side for the Cloud. Having an
an internal Cloud evangelist ensures that there is a balance of
evangelist protagonist and antagonist. In the culture
(supported by a of most large enterprises where change can
senior business be glacial by nature, find an energized
Cloud charismatic evangelist with business
champion)? acumen, give them the support they need
and let them do their thing.
⎕ 6.2.16 Have you The term lighthouse project refers to a
identified your model project that aims, besides its original
‘Lighthouse purpose, to have a signal effect for
Project? numerous follow-up projects as they look
towards it for inspiration and guidance.
Ideally choose something that everyone
can benefit from (e.g. payslips online?). In
addition to success, a great notoriety is
intended (Thanks Wikipedia).

62

..MS2019..
 The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.17 Do you Many today still completely miss the real
understand the value of public Cloud. When the National
concept of Institute of Standards and Technology first
‘above the IaaS defined Cloud in 2012 many disagreed with
line’ benefits of some elements of the definition (e.g. must
Public Cloud? include “orchestration”) because they
weren’t at that time widely available! An
adjusted temporary working definition
(read “IaaS”) was used for much of the
ensuing period. The legacy of the
confusion means that today many still see
Cloud and IaaS as synonyms. They are not.
The real value of public Cloud in the future
will come from the exploitation of the
PAYG 23 availability of Software, Platforms
and Functions as a Service (see Chapter 4).
⎕ 6.2.18 Are you The Cloud may deliver bottom line savings
in a position to and the point here is to ask if you are just
exploit these switching on Cloud or also mobilizing
‘above the IaaS people to work differently. To truly be
line’ benefits? Cloud native your organization needs
eventually to move away from concepts
such as servers and monolithic application
suites. The very essence of the server-based
model for how we deployed applications in
the past creates implicit scaling restrictions.
Ideas such as microservices, and the micro-
segmentation of applications has been
shaping the future of third-party Cloud
native application development for a while
now. If you have not expanded the edges
of your in-house IT environment (and
thinking) into the public Cloud, then
accessing, integrating and using these tools
to exploit your data and work differently
just isn’t viable.

23 Pay-As-You-Go
63

..MS2019..
The Business Questions > The Way Ahead

 Question Why this matters

⎕ 6.2.19 Have you The potential for runaway cost is one
begun to explore common objection inhibiting adoption of
mechanisms for Cloud. An important aspect of your
enabling (whilst production Cloud environment is ensuring
still governing) that cost transparency and management
the use of public reporting is inbuilt. To satisfy this demand
Cloud? there has been an explosion in the
availability of tools and specialized service
providers – use them.
⎕ 6.2.20 Do you If you only listen to vendors, you get a
engage vendor specific view. While there may be
independent legacy compatibility issues and loyalty of a
market analysts? long-standing relationship to consider, the
choices available in the Cloud are vast.
Specialized independent (and ideally sector
specific) market analysts can offer new
perspectives. If you take a moment to think
about the diversity between early adopters
and technology laggards in the Cloud world
it is more pronounced than in any
preceding iteration of IT change. The
Cloud savvy have accelerated away from
the pack. Accordingly, trying to catch up
with the leaders using only your own
resources is unlikely to work. Skilled
resources in this space are very sought after
– so get some help.
⎕ 6.2.21 Do you The #challenger is your internal secret
have your own weapon. Young, analytical, ‘tapped in’,
#challengers socially active and aspirational. By
platform? providing a platform for your #challengers
to bring forward ideas, and rewarding them
for doing so, is one way to keep a
#challenger mentality alive and attract
more to your firm.

64

..MS2019..
 The Business Questions > The Race is on

6.3 The Race is on

Ten years ago, few would have described technology as a race, and
it was only necessary to watch what your peers were doing and tune
in to what your preferred vendors were saying. Today that has that
all changed. With the economics of the Cloud and a dazzling array
of innovation being churned out fast, that is accessible not only to
your peers but also those plotting to disrupt your business. If that
feels uncomfortable – good, because if it did not then you would
be complacent and that is not where you want to be.
In the past a CEO would not usually be asked a question about
technology. If they duck the question today alarm bells start
ringing. If you read the news, then you will know the CEO carries
the can for technology – success and failure. Clearly, they are not
expected to deliver a sermon on Agile (Chapter 5). Rather they are
expected to articulate the big picture of how the organization will
respond to the opportunities of this new era of technology on tap.
If they choose to frame that with the aid of this book as their
Digital Operating Model, then they now have the way to do that.
Some questions follow, use them to go layers deeper in getting to a
known position that everyone on the leadership team is
comfortable with.

65

..MS2019..
The Business Questions > The Race is on

 Question Why this matters

⎕ 6.3.1 Do you Some organizations appoint a person or
wait for small team to go hunting for and be the
innovation to receptor (to suppliers) for absorbing
come knocking innovation. A detailed understanding of the
on your door? business is needed to recognize
opportunities to put new ways of working
to use enabled by technological innovation.
If digital is at the heart of every business,
then you will want that to be the very best
it can be. Do you expect innovation to find
you, or you to find it?
⎕ 6.3.2 Are you Sometimes we can’t see the wood for the
measuring the trees, and we follow old patterns because
right things, or that’s what we do. Data is the new science
are you just and art of business in the digital age. It is a
focused on major focus for the tech industry that
measuring things develops software to analyze data and for
right? customers hungry to analyze data. Don’t
recognize this? Perhaps you do not have
the expertise in-house - so then look
outside for assistance.
⎕ 6.3.3 Are you It may or may not matter? Every
in the race? organization will have some interest in how
they stack up against others. Are you
investing in technology or investing in
building a capability to use technology?
They are not the same. Few organizations
would say they do not have enough
technology, rather it is the case it is a heavy
load to carry. The Digital Operating Model
described in this book is about building a
capability to support the ‘business, people
and technology’ 24 for this digital age.

As reported in Microsoft document ‘Adopting the Microsoft Cloud Operating

24

Model’
66

..MS2019..
 The Business Questions > The Race is on

 Question Why this matters

⎕ 6.3.4 When Data is a strategic resource, an asset and
was the last time also a liability. We are witnessing increasing
the board was regulation around data privacy and
presented an ultimately the board is accountable. The
opportunity and board’s task is growing in complexity as it
risk assessment weighs opportunity and risk associated
of all data held with data. How do we take advantage of AI
by the while complying with our legal obligations
organization? to data privacy? Ethical questions arise that
need careful consideration and
communication for uniformity of
implementation. The new smarts are
building an organizational culture that sees
customers’ data as an asset for building
trust and not blindly applying technology
to analyze the data. Do that well and it is a
win-win.
⎕ 6.3.5 Are your Productivity is easy to define, often hard to
people still tied measure and an endless pursuit. The race
to the office? for a work/life balance, reducing the waste
(and environmental impact) and cost of
commuting time and convenience of
mobile working is wrapped up in the
productivity sum. The Cloud has been an
enabler of mobile working. Have you
explored all the opportunities available in
the Cloud in your pursuit of productivity?

67

..MS2019..
The Business Questions > The Race is on

 Question Why this matters

⎕ 6.3.6 Are your In an age where convenience is highly
business valued and aided by technology then throw
processes in some friendly. Making friendly can be as
customer and straightforward as making something
employee simple and reducing the work. “None of us
friendly? is as smart as all of us” 25. If you are
looking for innovation to apply to your
processes then the Cloud is the place to
look, that is where smart people
(developers) make available innovation.
⎕ 6.3.7 Do you If it ain’t broke, don’t fix it. True enough
say: “so what to yet in a fast-paced world the new skill is
transformation?” anticipating what’s coming down the road
that will knock you down. It is then a
question of how quickly you can react to
new situations be they threat or
opportunity. In this book we promote the
Cloud as the foundation for delivering
agility and building a competent
organizational practice around a Cloud
Adoption Framework gives you that
capability to respond quickly. The
bedfellow of anticipation is preparedness.
You are smart, you work out the rest.

25 Quote attributed to Kenneth H Blanchard

68

..MS2019..
 The Business Questions > The Race is on

 Question Why this matters

⎕ 6.3.8 Have you Right now, there is frenzy of new
done your technology matched by a frenzy of ideas as
homework? to how to put that to work. The oil in that
engine is data and a lot of that data was
obtained under consent. In some cases,
you are a custodian (Data Controller
and/or Data Processor) and not the owner
of that data. Before you throw data at new
technology e.g. Artificial Intelligence, check
your consent allows that and meets your
legal obligations, consult your Data
Protection Officer. You are familiar with
the responsibilities of a Data Protection
Officer, Data Controller and Data
Processor? If not, you have homework.
⎕ 6.3.9 Have you Whether you are a leader or follower it is
evaluated the valuable to have a position on technologies
opportunity and that have the potential to transform work.
threat of AI? AI is in this category. How would you
classify your current position? No position,
under review, quantified, applied? AI has
the potential to replace human work and
that has many implications to include how
it may impact your HR planning.

69

..MS2019..
The Business Questions > The Race is on

 Question Why this matters

⎕ 6.3.10 Have you The recent European Union General Data
evaluated the Protection Regulation (GDPR) has put
opportunity of data, and particularly Personal Identifiable
Big Data? Information (PII), in sharp focus. Before
you plough ahead with Big Data check
those involved understand the seven
principles 26 of GDPR and engage a Data
Protection Officer in case of any doubt. If
your tech team is not aware of these
principles (and many will not be) then they
could land the organization under risk of
investigation by a Regulator with potential
for significant financial and reputational
harm. Data privacy is not harmonized
around the world so education (of what is
permitted/lawful) must accompany the
evaluation of Big Data.
⎕ 6.3.11 Do you The promise of RPA is hotly contested
have a position with great differences of opinion about its
on the impact of impact and in particular the displacement
Robotics and of jobs. If you are an employer, then you
Process have a social responsibility to take that into
Automation consideration. Conversely, RPA could be
(RPA) in your used by competitors to undermine your
sector? business model and so it is wise to be alert
to this threat. Are there jobs that you have
difficulty filling or retaining staff? Is that a
safe place to expose RPA? Have you
considered the use of RPA to release
dependency on offshore processing? Big
decisions ahead.

26https://ico.org.uk/for-organisations/guide-to-the-general-data-protection-regulation-
gdpr/principles/
70

..MS2019..
 The Business Questions > The Race is on

 Question Why this matters

⎕ 6.3.12 Do you If you look back 5 and 10 years what has
have a vision for been the impact of technology on the work
your business in of the organization? Probably you will
a digital age? identify that technology has had a big
impact and more recently the Cloud will
have driven that agenda. So, in your vision
what are digital capabilities that your
organization must excel at and why are
they mission critical? 27

27A mission critical factor of a system is any factor (component, equipment,

personnel, process, procedure, software, etc.) that is essential to business
operation or to an organization. Failure or disruption of mission critical factors
will result in serious impact on business operations or upon an organization, and
even can cause social turmoil and catastrophes. Source: Wikipedia
71

..MS2019..
The Business Questions > Agile – small word big work

6.4 Agile – small word big work

If you want to be that organization that innovates and delivers a
true digital experience for your customers and employees, you need
to become more of a software company.
As this book set out to explain, the Digital Operating Model is
where people and technology huddle together to put that into
effect. How?
Agile is well established in the software engineering space, but
increasingly is moving into every aspect of service creation across
the business and IT. Now we turn your attention to questions that
you will want to discuss with colleagues.
Agile raises questions: Do we pass? Can we take that chance? Are
we ready for Agile? If not, what do we need to do to be ready?

72

..MS2019..
 The Business Questions > Agile – small word big work

 Question Why this matters

⎕ 6.4.1 Do you If you do employ software developers, then
employ software don’t stop there. They should be given rein
developers and to focus some of their time working closely
data analysts? with the business teams. When they truly
understand the needs of the business team,
they are better equipped to deliver a
solution. Different? Better? Good idea?
You won’t know until you put it into
practice.
⎕ 6.4.2 Have you If the business and IT are to begin to
invested in Agile engage in iterative ongoing Agile
training – and development and innovation, then a good
not just within starting point is to make sure everyone
the IT team? understands the big picture and reasoning
behind such an approach. They also need
to learn the language. How many of your
business leaders know what a “sprint” is?
Chapter 5 was written for this very
purpose.
⎕ 6.4.3 Do you NASA and “Big Pharma” don’t get it right
have an first time every time. They accept the
experimentation limitations of and learning from investment
budget? in experimentation. The IT and Business
Team staff need to have some flexibility to
try (and fail fast) at new ideas, and to do so
they need a “slush fund” with which to
experiment. How else will you sow the
seeds of innovation?
⎕ 6.4.4 Do you Such a measurement will often reveal a
measure the surprisingly slow (glacial even) innovation
average time it cycle. You may think you’re reacting
takes for your quickly, but if you take all of the steps in
organization to the process into account you may be
move from idea unpleasantly surprised. The issue most
to delivery? commonly cited as an obstacle to adopting
Agile is culture. Who will take the lead?

73

..MS2019..
The Business Questions > Agile – small word big work

 Question Why this matters

⎕ 6.4.5 Do you Customer satisfaction surveys are
constantly test omnipotent, yet customers can be fickle. If
your your surveys show a change in how your
understanding of customers perceive your products/services,
what matters to how quickly can you react? In this digital
your customers? age with instant and continuous drip feed
of feedback it sets the challenge to
respond. Even if you think you have tuned
your customer experience to be the best it
can possibly be, you should never stop
teasing and testing your marketplace. The
adoption of a radical idea that changes an
entire sector can often simply be a factor of
target market maturity or timing. You
should begin to adjust your mindset to be
as happy with a failed result as with a
successful one. Failure can teach you as
much as a success. Take a look at your
email and voice-based call center to service
younger people...they don’t like to use
either!
⎕ 6.4.6 Have you Rather than being mandated top down,
enabled cross- Agile is often best embraced by an
infection of organization through the process of cross-
enthusiasm? infection of enthusiasm. Pick a high
profile, low risk project – such as that time
reporting system that everyone loathes –
and use a process of iterative feedback and
improvement involving everyone to make
it sing. Use staff voting on ideas to drive
the outcome. This will allow everyone to
see and participate in a successful example
of Agile at work and let them understand
how it can benefit their customers’
experiences too. They will be lining up to
drive their own project next.

74

..MS2019..
 The Business Questions > Agile – small word big work

 Question Why this matters

⎕ 6.4.7 What is When you’re looking to do something new
your mechanism like launching a new client service, the
to budget for business you will typically look to IT to
and deliver new design, quote and deliver the project.
services? Individual projects and the budgeting
associated with this can cause excessive
delays and require detailed functional
specifications which can change rapidly in
this fast-paced digital world. You need to
consider alternative more flexible
mechanisms for interfacing with (and
budgeting for) IT enablement.
⎕ 6.4.8 Have you One way to avoid excessive governance
considered and overhead on innovation is to establish
having an teams of digital specialists whose sole job is
always-on to work through a backlog of innovation as
innovation team? imagined by the business and your clients.
There is a fixed, budgeted cost for this
team per month and its efficiency can be
measured via the velocity of new
capabilities (or problem fixes) they are
delivering.
⎕ 6.4.9 How One of the advantages of having projects
would we and budgets is to help track the success of
manage and these projects and avoid “scope creep”.
measure Success here is, however, often measured
innovation in the number of reports and artefacts
teams? created. But isn’t “scope creep” really just a
result of people trying to get what they
actually want? Moving to a more Agile
model of innovation management can have
significant advantages, such as measuring
progress in terms of value delivered to the
business.

75

..MS2019..
The Business Questions > Agile – small word big work

 Question Why this matters

⎕ 6.4.10 How Agile in its raw form is good at getting
would we report things done, but poor on reporting what
on this has and has not been done. “We’ll keep
innovation? going and let you know when we’re done”
isn’t very helpful. Frameworks such as
SAFe® and Microsoft’s Cloud Adoption
Framework can provide the required
structure, business alignment and reporting
that traditional project management
methodologies can’t really deliver in an
agile world.

76

..MS2019..
Part 2 Technical

Thinking of

Building a Digital Operating

Model with the Microsoft Cloud
Adoption Framework for Azure?

Ask the Smart Questions

'What the technical team need to know'

77

..MS2019..
78

..MS2019..
 A new operating model

Chapter

7
7A new operating model
The definition of insanity is doing the same things over and over again expecting
a different result
Falsely attributed to Albert Einstein (Theoretical Physicist, 1879 – 1954)

Why do we need a new operating model?

I N recent years, IT has been viewed through a lens of cost

reduction. A successful CIO or CTO has been one who tamed
the flow of cash exiting an IT department’s shrinking budget.
They would attempt to tune an already well-oiled machine to
extract additional value and productivity.
Processes were brought in to streamline and control every IT
operation.
Information Technology Infrastructure Library (ITIL) emerged as
the dominant viewpoint of an efficient IT operation. ITIL included
Service Design, Service Transition and Service Operation. These
were applied to Availability, Configuration, Testing, Release,
Knowledge, Change, Event, Incident and many others. It became,
and is still viewed now, as the epitome of what good looks like.
As we read in Part 1 there are two competing forces in any
business or IT operating model: agility and control. This is so
whatever the shape or size of the organization. The business
operating model craves agility whereas the IT operating model
strives to deliver control.
The trend of the last two decades has been toward outsourcing,
and this has exacerbated the traditional divide. The business yearns
79

..MS2019..
A new operating model

for new capabilities, services and business models, while its IT

partner looks to maximize profitability by resisting change and
increasing control. The basis of outsourcing is that an outsourcer
can deliver an IT service much more cheaply than an internal team.
That is because it can offer specialization and economies of scale.
But it also requires a slick and controlled operating model. Such a
model penalizes (through change control) doing anything
differently from the way it is done today.
Change control is frustrating for a business having a traditional IT
operating model: ‘Why are we being penalized for wanting to do
something better than the way we do it today? Especially if our
proposed change increases customer satisfaction and delivers new
services?’ Often for the business ‘changing something means they
have found a better way of doing something or need IT to deliver
services in a different way.’
The IT provider complains that the business just doesn’t get it.
Change is hard. Change is difficult. Change is risky. Change is
dangerous. Change is something to shy away from and discourage.
In this book we challenge these preconceptions. We also challenge
the view held in pretty much every organization on the planet that
change is a necessary evil.
We propose a new model. We want to construct a new relationship
between business and IT. This will be a new grand bargain where
IT lives at the center of an organization’s business operating model.
IT will sit at the top table in determining strategy. IT will feel
empowered and be valued in the drive to differentiate the
organization’s marketplace.
To do this though, we the IT department need to behave
differently. We need to be equal partners rather than just service
providers. We must understand the context of the requests we
receive. We should become more proactive in making
recommendations, suggestions and proposing new models. We
must be more deeply engrained in the business operating model.
At the same time, the business also must behave differently. It
needs to stop thinking of IT as a support function, detached from
the important work of actually running the business. IT should not
be put in a box – a service to be provided as cheaply and efficiently
as possible. It should definitely not be viewed as a nuisance getting
80

..MS2019..
 A new operating model

in the way of making more money. The business should think of IT

as a strategic advantage, able to bring about digital differentiation
within an increasingly cut-throat market.
When the entire organization decides to give customers and
employees the best experience, it develops a new attitude toward
change. For both actors, change becomes something to embrace
and to invest in. It becomes something desirable rather than
something to shy away from or discourage.
Once this decision is made, the question becomes one of “how” to
move forward. This is where the Microsoft Cloud Adoption
Framework for Azure comes into play.

The changing roles of CIO, CTO and CDO

The CIO role has changed over the years. This has been in
response to different requirements for the management of
computing systems and data. In particular, the Chief
INFORMATION Officer was responsible for all of the
organization's data and was its chief data custodian. But that didn’t
last. As we approached the Y2K years, the role of the CIO
concentrated more on financial control and long-term planning for
IT systems. In particular, the CIO was expected to forecast the
computing needs of an organization three years ahead. They were
expected to do this with insufficient data about how changes in the
world would affect the organization’s IT capability and spend.
When the Cloud PAYG (Pay-As-You-Go) model came into being,
the power of the Chief Technology Officer increased. More
importance was placed on making the correct choice in the short
term as well as the long term. Less weight came to be placed on the
cost of IT and more on how organizations could get a ROI from
their IT spend. The Chief Technology Office and the CFO then
teamed up, and we saw the decline of the CIO role.
In the last five years the requirements have changed again. Machine
learning and AI services have become available to drive greater
insight into an organization's data. These services are cheap and
effective. Now, with the world of data ever changing and the
growing importance of data for the organization, the role of the
CIO has emerged once more. But interestingly enough, it has re-
emerged in the form of the CDO (Chief Data or Digital Officer).
81

..MS2019..
A new operating model

From an outsider's perspective, it’s important to make the

distinction. The CIO is still looking at Cloud from a monetary
perspective. This includes the operating model that comes with
data and next generation workloads. The CTO cares about the
adoption of innovation and the ROI on technology adoption. An
operating model that can best support this is an operating model
that provides guard rails to support a safe environment in which to
succeed, or fail, both culturally and technically. Finally, the CDO
role is important. It is the first C Level role tasked with driving the
adoption of public and in some cases multi-Cloud. This takes
advantage of the power of AI from one Cloud while leaving
standard workload management to another one.

How are Cloud Operating Models and the

Digital Operating Model different?
In this book we refer to two operating models. The overarching
model that covers both Business and IT we refer to as the Digital
Operating Model. A Digital Operating Model describes how the
business and IT can better interact with each other at a macro level
within a given organization. The specific operating models that
govern the lifecycle of a given IT service line is described as a
Cloud Operating Model. A specific Cloud Operating Model (COM)
has many similarities with a traditional IT operating model
(ITOM). The fundamental difference concerns how change is dealt
with and the high levels of automation possible with Public Cloud
services. In a Digital Operating Model, IT continues to deliver
everything it has delivered in the past such as stability and
availability. But this continues in a world built on change – a world
where everything changes all the time.
A traditional IT operating model starts from the pretense that
change is difficult and dangerous. There is the fear that things often
break when you change them.
But what if we could build a model where change is not dangerous
– or at least is orders of magnitude less dangerous? That model is
what we will try and construct in the following chapters.

82

..MS2019..
 Strategy and service providers

Chapter

8
8Strategy and service providers
Simplicity is the ultimate sophistication
Leonardo da Vinci (Polymath, 1452 – 1519)

Setting off in the right direction

T HE first step on a journey to a Digital Operating Model is

building a Cloud strategy. A move to the Cloud is a big deal.
It’s a big decision – a decision that will permeate every part
of your business and impact every aspect of your existing IT
operations. It can also be risky and is not something to do lightly.
Building this strategy can be involved. You will need to engage a lot
of people across your business. Everyone must buy in to the
decision that this is something we want to do. If they are not
entirely convinced and not fully committed, we must know to what
extent they are in support.
You might canvass your business to make a case for moving to the
Cloud. The consensus might be negative or lukewarm. They might
say that it’s just too difficult: ‘Our IT is too elderly. We can’t move
our data to the Cloud. We are so super special, and we just can’t do
it.’
We understand. Cloud is not right for everyone. We wish you all
the best. Maybe we’ll speak again in the future.
Nearly all organizations react in this sort of way to begin with. That
includes the ones that have since fully embraced Cloud. But to be
successful you will need to secure buy-in from the business and
from the key IT stakeholders. As Friedrich Nietzsche once said,
83

..MS2019..
Strategy and service providers

“He who has a why to live for can bear almost any how.” The same is true
for businesses. Without a why this can be a very difficult path to
strike out on. We will cover this topic in a little more detail later in
this chapter when we introduce the Microsoft Cloud Adoption
Framework section on “Define Strategy”.
If you want to jump straight to it and read about Strategy within the Cloud
Adoption Framework for Azure, visit: https://aka.ms/adopt/strategy

For those still following, let’s get down to business. We’ll assume
you want to do this and so it’s not if, it’s how. We’ll assume you
landed on something like “Cloud First”. This advocates Cloud
wherever possible, unless there’s a really good reason for not
adopting it.
Once you’ve got the green light, we’ll need to get down to the
detail. What does Cloud-first mean? How do we build a strategy
around this? This book is not designed to educate the reader on the
NIST definition of Cloud computing and does not get drawn into a
debate about what is and what isn’t really Cloud. When we talk
about Cloud in this book, we are talking predominantly about
public Cloud or Hyperscale Cloud. These are the providers who
invented the concept, not those who rebranded their existing
offerings.

Figure 7 - Guiding principles for IT Strategy

84

..MS2019..
 Strategy and service providers

A true Cloud strategy should focus on public Cloud. This will

enable you to harness the innovation we spoke about in Part 1.
Public Cloud is required if you really want to digitally transform
your organization.
We now need to break Cloud down into the three main groups:
IaaS, PaaS and SaaS. Again, this book will not dwell on the
definitions of these things. We assume the reader understands these
concepts at a high level. Your strategy needs to be cognizant of the
differences between these service models and address each
separately.

Cloud Center of Excellence

A great place to start on a journey towards building these standards
and service-centric patterns is to create a Cloud Centre of
Excellence.
Within a Cloud Centre of Excellence, teams come together to build
the guiding principles. This team does not need to be a standing
team of permanent members but should represent business and
service staff. The objective of the CCoE is to strike a balance
between speed and stability. It is a mechanism to move an
organization from a regime of control and central responsibility to
one of freedom and delegated responsibility.

Multi-Cloud
The two main building block services for your core infrastructure
will be IaaS and PaaS – server hosting and application hosting.
According to Gartner (https://azure.microsoft.com/en-
gb/resources/gartner-iaas-magic-quadrant/en-gb), there are not many true
Cloud infrastructure providers to choose from. There are Amazon,
Microsoft, Google, Alibaba, Oracle and IBM. Assuming you don’t
have significant operations in China and that you’re not locked in
either to Oracle or IBM, only three are left.
Your next decision is: Do I as a business support, one, two or all
three of these providers?
What does all this mean? There is multi-Cloud and there is Hybrid
Cloud. Some very large organizations may use both.

85

..MS2019..
Strategy and service providers

Multi-Cloud is a much-debated topic now. Some in the industry

define the concept as a common technology layer hosted by
multiple vendors. Others define it more as a deployment model
that requires a common abstraction layer between different
environments.
We will treat multi-Cloud as something that requires the
management of different Cloud service providers. They will be
both public and private and use different tools and services in a
common way.
Suppose you decide to provide a suite of IaaS services from more
than one Cloud provider. You will need to include the tooling and
process so that services deployed into one Cloud meet the same
configuration and governance standards as another.
Even for something as simple as a single virtual machine host, you
may have to replicate effort in the management and maintenance of
service blueprints. Management tooling may be required in multiple
locations. Your brokering and orchestration layer must support
multiple methods of addressing the Cloud provider's fabric. This
applies whether the layer is highly automated or invoked on
demand by staff. The fabric is mostly bespoke for each vendor.
The tools required for service management also need to consider
the subtleties of each vendor’s Cloud service environment.
In the Microsoft Cloud Adoption Framework, the Ready phase
covers this topic in the context of Azure based workloads. The
Ready phase talks of Landing Zone considerations and best
practices that provide guidance on how to address a series of
design principles. These principles apply equally to all Cloud
providers but may be solved in different ways.
There are, of course, ways to abstract this complexity. This can be
done by consuming lower level services such as containers. But
most heritage IT estates are a long way from being able to move
more than a small portion of services to benefit from this type of
technology.
There are several vendors and platforms which provide this
abstraction for you, but they can be complex and expensive. They
also often fail to take advantage of niche platform-specific
capabilities from other vendors, such as Azure Cognitive Services.

86

..MS2019..
 Strategy and service providers

There are good reasons for adopting multi-Cloud. But there are
also bad ones.
For example, a good reason would be a scenario in which you
consume services that are suitable for a certain public Cloud
provider, while having other services that lend themselves to a
different provider. If you are in this situation you may be forced
into a multi-Cloud world by your application layer.
In this case you may find yourself consuming SAP S4/HANA
public Cloud as well as Oracle, IBM Bluemix, and Azure.
Alternatively, a poor reason for choosing a multi-Cloud approach
would be the fear of vendor lock-in. The ability to move workloads
between vendors to get the best price is not a good criterion.
We will argue in this book that multi-Cloud entails greater
management complexity and operational risk. Unless dictated by
regulatory bodies, this potentially outweighs any risk of committing
to a single Cloud provider. Also, many Cloud providers make a
charge for offloading data from their services. An architecture that
moves vast amounts of data between Cloud providers doesn’t
make sense.
Instead, we encourage organizations to think about general-
purpose and specialized operating models. We encourage
organizations to pick one Cloud (we recommend Azure, naturally)
as their primary, general-purpose Cloud. Your general-purpose
Cloud should deliver end-to-end capabilities, including your control
pane. It should deliver your core management and monitoring
services across all Clouds. Any services which are common and
non-differentiated between vendors (such as VMs) should only be
supported within your primary Cloud. If you believe another Cloud
provides a service which doesn’t exist within or is substantially
better than the one within your primary Cloud, support just this.
Your specialized operating models can therefore be much simpler
and more focused. You get the benefits of multi-Cloud without
having to build multiple general-purpose operating models.

Hybrid cloud.
Hybrid Cloud is an easier concept to grasp. It relates to the location
of services within a management boundary.

87

..MS2019..
Strategy and service providers

Hybrid has been a core part of Microsoft’s narrative since

Windows Server 2012 R2, its first Cloud OS. The idea is that if you
were running Hyper V and System Center across your on-premises
estate and were consuming Azure public Cloud services, you were
in effect within a single management plane. You were then able to
shift workloads around to meet whatever business and technical
circumstances you were faced with.
This conversation has matured over the years with the introduction
of technologies like Azure Stack. This allows for a near seamless
operating model across multiple physical hosting options: on
premises, partner hosted and public Cloud.
For those who cannot afford Stack, you can still get a high level of
integration out of the box between your on-premises environment,
a third-party hosting provider, and Azure.
Microsoft also provides tools like Azure Backup, Azure Site
Recovery and Azure Monitor. These will run both across on
premises and Azure, providing a unified management plane for
these operational services.
Hybrid Cloud can, therefore, achieve a location-based paradigm
within a single set of management tools. If done right. If done
wrong, hybrid can be a complexity masqueraded as flexibility.
We have now defined the concepts of multi- and hybrid Cloud. We
can move on to a closer look at the architectural concepts that will
determine which model you adopt.

Your estate
IT estates can be divided into end user services, which we describe
as “Intelligent Workplace”, and back office infrastructure and
application services which we call “Intelligent Cloud”. The strategy
you adopt will be determined by the capability and direction of
travel for these services across domains.
Your new end user compute services will incorporate public Cloud.
This has an impact on the architecture of your back office and how
you deploy applications to it.
Bimodal IT is a key aspect of the new IT world you are planning.

88

..MS2019..
 Strategy and service providers

Bimodal IT
Gartner’s describes a two-speed model of IT deliver.
(https://www.gartner.com/it-glossary/bimodal/). In this model heritage
IT assets exist alongside newer, more agile Cloud-native capabilities
and service lines.

Figure 8 – Bimodal IT© Gartner Inc

The Bimodal IT concept is important. We will elaborate later in the

book. At this stage you should understand the core differences
between two modes. The closer you get to Mode 2, the more
valuable the IT organization will be to the business.
The primary focus of the two modes is summed up in their
respective goals. The goal of Mode 1 is reliability. This means up-
time and stability. This mode is geared to fight change. Its focus is
the efficiency of the IT estate.
Mode 2 by contrast is all about change. It describes change in a
positive light; change implements new ideas quickly and favors
business agility.
You will be developing an IT strategy to incorporate public Cloud
services. As you do so you need to think about how you architect
services to best take advantage of the fast pace of change in public
Cloud.
Your IT organization needs greater Cloud maturity. This is because
it will allow your services to respond quicker to the fast pace of
change in the business.
89

..MS2019..
Strategy and service providers

Figure 9 - Cloud Maturity Model

The ideal is that your businesses should be able to realize revenue

from an idea by quickly enabling the required IT services to deliver
it.
These IT services should be designed in such a way that they can
be spun up and down on demand. The maintenance and health of
these services should be highly automated.
You may not achieve all this at the first pass. But the closer you get
to the optimized model the greater will be the benefit to you and
your businesses.
One of the keys to success is to operate both modes initially. This
is why it is called Bimodal IT. You can’t just abandon your Mode 1
heritage workloads. The estate is running your business today and
much of it will be running your business in 12 to 18 months from
now.
It is also important that you begin to incorporate services
architected for native public Cloud.
Many organizations begin by introducing small applications and
services that supplement existing, deeply embedded COTS
applications.
An example would be extracting key data from an on-premises line
of business application into Azure Data Services. You could then
write a mobile app to consume this data. The app might serve a
single business function and might be read-only. You could create a
business process to update data in the underlying application.
When the data got updated, you could automatically push this
update to the app and notify the user. This model allows users to
see this new piece of information in near real time without you
having to rebuild or re-architect the underlying legacy line of
business application.
90

..MS2019..
 Strategy and service providers

A small set of applications like this can spring up around a core

LOB application. They will add huge amounts of value. There is no
need to invest large amounts of effort and money in re-engineering
the core application at this stage. This composite model of
application development is particularly useful as you embark on
your journey to the public Cloud.
In the future, major software vendors will rewrite their applications
to work natively in the public Cloud. Until then you will need to
support your IT in both modes.

The Microsoft Cloud Adoption

Framework for Azure – Define Strategy
Having discussed some high-level considerations, let’s now jump
into the specifics as they relate to Microsoft’s view on strategy. As
you would expect, Microsoft kicks off their Cloud Adoption
Framework for Azure with a stage called “Define Strategy”.
To read more about Strategy within the Cloud Adoption Framework for Azure,
visit: https://aka.ms/adopt/strategy

In this phase, the framework speaks to a need to understand the

motivations of a Cloud migration. It also advises to understand the
justification and outcomes from a business perspective of any
migration program and to identify a potential first workload or
workloads to move.
A recent discovery is organizations that understand their
motivation are far more likely to undertake and complete a Cloud
transformation program.
When assembling your list of potential business outcomes, the
Microsoft Cloud Adoption Framework provides some categories
for this discussion. It organizes this thinking into a business
outcome framework that covers subjects such as fiscal, agility and
performance outcomes. There is also some guidance on how to
capture this information.
This step is critical as you commence your journey to a Digital
Operating Model. Unifying the business and technology teams
around a common strategy and set of goals is paramount.
Many Microsoft partners will have consulting engagements in this
space. You will need to cover the subjects of aligning the business
91

..MS2019..
Strategy and service providers

strategy to the IT strategy, the readiness of the environment to

connect Cloud services, the relative maturity of your organization
across a wide spectrum of elements, and finally the economic
position pre- and post-Cloud migration.
The Cloud Adoption Framework Benchmark tool provides an
excellent way to understand the current state of relationships
between your business and technical teams. It also allows you to
rate your relative Cloud maturity with peers in the market. Visit:
https://aka.ms/adopt/gov/assess
It is also essential to have a business case that stacks up. Your
driver might be a datacenter exit and the avoidance of a large
capital expenditure at one extreme. Alternatively, you might be
motivated by increasing the speed of the IT organization and
reduce your “time to value”. In either situation, it is important that
you know what the 3–5 year likely cost profile is.

Service Provider Relationship

There is another important aspect of adopting public Cloud. There
will be a change in the relationship between you and your Cloud
vendor.
In the past you bought software from, say, Microsoft, Oracle, or
IBM. You either ran this software yourself or paid someone else to
run it for you. This is a classic ITOM model. Responsibility for the
software is separate from the responsibility for delivering it.
In a Cloud world, these functions are often blurred. The creator of
the software now delivers the software (or platform) to you as a
service.
Occasionally in a non-Cloud world, actions of the software vendor
can cause you real pain. A bad update for example. But most of the
time you are responsible for and in control of your IT service.
In the present world, when something goes down, anyone, even
the CEO, can come down and get grumpy with you until you fix
the problem. Their presence doesn’t make the situation any better,
and in fact often makes it worse. But at least they have the
satisfaction of feeling that they are getting involved and, in their
eyes, helping to get the problem solved.

92

..MS2019..
 Strategy and service providers

In the new world of Cloud, there is no room for them to come

down to! No one to come and shout at! There is nothing anyone
can do to assist in bringing the service back up. There’s only one
thing you can do: that is to let the users know there’s a problem
and that it’s in hand. Then you can sit on your hands and wait. This
can be incredibly disconcerting for both you and the business.
Suddenly, it no longer matters who shouts loudest. Services will be
brought back online, as quickly as possible, in the best technical
way. If SLAs have been breached, you’ll get credits back. There are
no consequential losses, and no punitive damages.
Whilst this can be off-putting and indeed scary in some situations,
there are nevertheless some positives. When things go down, it’s
stressful – very stressful. We’ve all been there late at night or over a
weekend battling to get something back online. It’s not something
you’d wish on your worst enemy. It’s not something you’ll miss
from your past.
In the Cloud world, you may not feel personally as important as
you used to be. You’re actually part of something much more
important. You are part of a much bigger collective. You’re just
one of millions of users who are also getting very grumpy. There’s
nothing like a service that caters to millions of users for getting the
attention of vendors. And so, on the very rare occasions that things
do go down, they tend to come back up very quickly.
You can, of course, mitigate some of these risks with a multi-Cloud
and multi-vendor approach. But you need to carefully consider the
pros and cons of such an approach. The question to ask is: Is some
downtime acceptable if it has significant savings in terms of cost
and complexity? Do you really have to push for that 99.999 percent
SLA?
This change of relationship is a profound one. It is a change from a
supplier/customer relationship to a partnership relationship. As
you move to the Cloud, your fate and your service provider’s fate
become intertwined. Their success is your success; their failure is
your failure. It’s a momentous change, and one you need to be fully
prepared to accept. The buck no longer stops with you; it stops
with them as well as you. You need to be immersed in this idea of
shared responsibility.
This is a new normal approach, and one to be shared by service
providers and consultancies.
93

..MS2019..
Strategy and service providers

Service Level Agreements

This new partnership-based relationship necessitates KPIs and
other measures to govern it. The service descriptions and operating
commercial agreements now become enormously important
documents. They govern the gives and gets of the relationship
down to the lowest level of detail. The service level agreement is
the more important one. It commits providers to make available
xyz service to you, a minimum of 99.9x percent of the time.
For details of the SLAs provided by Microsoft Azure, visit:
https://azure.microsoft.com/en-us/support/legal/sla/

Your partnership depends on this key metric. It will govern what

success looks like for your provider. As a guide, 99.9 percent
service corresponds to about 45 minutes of downtime a month. At
99.99 percent that figure is reduced to about 4½ minutes a month.
Providing service downtime doesn’t exceed those limits; all will be
well.
However, it’s important to dig deeper into these SLAs. Different
services have different SLAs. Office 365 guarantees 99.9 percent
uptime. Single instance VMs in Azure come with a 99.9 percent
uptime guarantee. But this can be increased to 99.95 percent with
an availability set, and even to 99.99 percent with two availability
sets. Just to increase the complexity, SLAs can interact with each
other. Suppose you have a load balancer with a 99.95 percent SLA
in front of a VM with a 99.95 percent SLA. This is equivalent to an
SLA of only 99.9 percent (99.95 percent * 99.95 percent = 99.9
percent). You depend on both and each can go down .05 percent
of the time. That gives a combined downtime of 0.1 percent. It’s
vital to understand these compounding SLAs. At the least, you
should at least understand the concept. It is highly relevant to your
evaluation of what a service provider is committing to.
These are some of the issues to guide you in selecting your multi-
region or multi-Cloud strategy.

Support
Support is a broad concept and an issue we will investigate in more
detail in the section on service management. But for now, it needs
94

..MS2019..
 Strategy and service providers

to be pointed out that it is vitally important and should be planned

for up front.
The questions to ask are: When something goes wrong, who is
going to respond? How are they going to interact with the service
provider? What kinds of support does your service provider offer?
Are there escalated tiers of support available either to you or to
other providers you partner with?
Getting the support relationship right and defining it properly are
keys to the success of your new Cloud environment.
We cover many of these topics in more detail in other chapters.
For the purpose of this strategy chapter there are some high-level
concepts to consider.
The Microsoft Cloud Adoption Framework covers this subject
matter in the “Plan” phase. The key sections here to familiarize
yourself with are the section on “Initial Organization Alignment”
and the section on Skills Readiness.
Firstly, there should be a single owner for your service management
layer. All the responsibility for governance and service operations
need to be given to someone with a holistic view of your world.
This person must also understand the business. It is possible to
outsource this responsibility. But there are strong arguments for
keeping it in house.
The orchestration of all IT services to the business requires an
impartial overseer. Service models become ever more complex.
There can be multiple vendors and partner relationships. This
overseer primarily represents the business and his or her job is to
keep everyone honest.
Secondly, you need to own your IT strategy. Bringing in expertise
from external advisors and consultants is a good thing. But the
Enterprise Architecture function and the ultimate arbiter of the 1-,
3- and 5-year plans should be an internal resource. We touch a bit
later on the Scaled Agile Framework (SAFe®) which has its own
Portfolio function that is a good replacement for TOGAF.
This Enterprise Architect or Portfolio Lead manages the direction
of travel for the organization’s IT services. As such they need to be
accountable to the CIO and therefore to the business itself.

95

..MS2019..
Strategy and service providers

The EA / office of the CTO should also be thinking of ways to

open new revenue streams for the business. This will be done by
monetizing data and services where possible.
Lastly, you should build a cohort of suppliers behind the scenes
consisting of specialist best of breed partners. You will also need
partners who can provide scale when it is needed.
Many of the Global SI community are struggling to convert to
Cloud services as these services cannibalize existing revenue.
However, they are excellent at providing scale services such as
service desk and application support.
There are lots of new Cloud-focused MSPs that can be
incorporated into your service model. They will sit between you
and the services and vendor management of your chosen Cloud
provider.

96

..MS2019..
 Procurement and financial governance

Chapter

9
9Procurement and financial
governance
The way to stop financial joy-riding is to arrest the chauffeur, not the
automobile
Thomas Woodrow Wilson (United States President, 1856 – 1924)

Building a new procurement model

T HE first step on a journey to a Digital Operating Model is

to build a model to support procuring your new Cloud
services.
Procurement and financial governance are completely different in a
Cloud world. Instead of purchasing assets such as hardware or IP
rights, we consume various kinds of services. We are buying levels
of capability and offerings based on very different pricing models.
To understand how to procure and govern these new Cloud
services you must understand what it is you are buying. Different
metrics and measures apply to each service.
Some of these models are based on a fixed price per unit or per
month. Typically, that applies to SaaS services. But many
components have variable pricing – PaaS and IaaS services are
examples.
It’s important, as you build your procurement and financial
controls, to take account of variability. As a rule, it's impossible to
predict how much you will be spending on your Cloud services
each month. Instead, think about a model which includes pre-
97

..MS2019..
Procurement and financial governance

defined, and pre-authorized blocks of spending. Additional

governance and controls can be invoked if these amounts are
exceeded.
The Microsoft Cloud Adoption Framework covers a section which
Microsoft refer to as “What is Cloud accounting?” in the business
justification section of the Strategy phase. This section covers
concepts like Showback and Chargeback and combines with a
section called “Creating a financial model for Cloud
transformation” to give you the tools you need to begin to unpack
this topic.
To read more about Cloud Accounting, visit:
https://aka.ms/adopt/strategy/cloudaccounting

The next question is: From where are you going to procure these
Cloud services? In the past you might have bought hardware and
software from a reseller or distributor. In the case of Microsoft
licensing, you might have bought from a Licensing Service
Provider (LSP). This could have been part of an Enterprise
Agreement.
These licensing vehicles and partners play less of a role in the new
Cloud world. Instead, organizations such as Microsoft have
overhauled their licensing programs. They have adopted a more
flexible, consumption-based model.
In future you will likely purchase licenses and Cloud capacity from
Cloud Solutions Providers (CSPs). They handle both licensing and
professional and managed services. These partners may source
these licenses and capacity directly from Microsoft (direct CSPs).
Alternatively, they will get them from distributors (indirect CSPs)
before selling them on to you.
One of the advantages of a move to a Cloud-based consumption
model is clarity and transparency of pricing. Likewise, one of the
disadvantages of a move to a Cloud-based consumption model is
clarity and transparency! The price is the price which is the price. In
the old world, RRPs were largely ignored as organizations cut
individual deals with vendors. There were large discounts available
in line with scale. The best time to buy Microsoft licensing was in
June, at the end of Microsoft’s financial year.
In the Cloud world there are a few discounts for the very largest
customers – those who spend millions of pounds per year. But
98

..MS2019..
 Procurement and financial governance

discounts and complex pricing models are no longer available to

most customers. Simplicity is, in the case of Cloud, the ultimate
sophistication.

Procurement pipeline
At the start of your journey into Cloud you must identify your
strategic Cloud provider. After that you need a plan to bring the
services on board. Typically, these onboarding services are not
delivered by the Cloud provider themselves but are undertaken by
specialist partners.
Many organizations choose Microsoft as their strategic partner.
Since you are reading this book, presumably you are one of them.
Congratulations, you made a good choice. You’re comfortable that
Microsoft has ticked your various boxes. You’re good to go on
Office 365 and/or Azure. It's time to crack on. Don’t let a
procurement pipeline get in the way of starting to use that right
now.
There will typically be other SaaS solutions which you need to
procure. You need to build a process for this procurement. The
process is not dissimilar to that for identifying and procuring
traditional software. You’ll need to go to market with your
requirements and see who comes back with what. You might want
a specialist solution for the automation of business processes. Or
you might want an intelligent chatbot solution.
Whichever solutions you choose, it is important that they sit within
the realm of your Digital Operating Model. The questions you will
ask of a SaaS vendor will be different to the ones you will have
asked when procuring anon-premises solution. You need to ensure
that the platform that their solution is built on will integrate with
your environment. If you want to run it as a service, without your
environment, it must be compatible with Azure. You need to
understand how it operates and how you and the vendor will
support it.
Getting this procurement right will ensure you can support
requests for SaaS solutions from the business. The SaaS solutions
must fit their requirements. They must also meet your availability
requirements. Finally, they must be secure to your standards and
work for you.
99

..MS2019..
Procurement and financial governance

Financial Operation (FinOps)

Procuring Cloud services is difficult. This is because of the
variability in the commercial models of different providers and
services. To deal with this the concept of FinOps or Financial
Operations has been introduced. This is a new kind of operational
function within your organization responsible for procurement and
billing of Cloud services.
To horribly misquote Churchill: Never in the history of IT has the
power to spend so much rested in the hands of so few.
Every configuration change in Cloud has a financial consequence.
These come into effect immediately a service is live and ready to
use; likewise, immediately after it's shut down or resized. This is
not a bad thing. Indeed, part of the benefit of public Cloud lies in
your power to manage the cost and benefit of IT services. This can
be done almost in real time. At the same time, these costs become
more transparent to the business.
The more traffic you send to a service the more resource it can be
allowed to consume.
Equally, as demand for a service goes down, they can have
resources removed. This will reduce the run cost.
An unfortunate example of this can be seen in the following story.
An early customer of Azure engaged a consultant to undertake an
introduction to Azure workloads for their company. The
consultant got to the HPC (High Performance Computing) services
topic just before lunch. He kicked off the provisioning of a 50-
node cluster as everyone left to eat. When they all got back 45
minutes later the environment had been provisioned. But it had
burned through his free £120 MSDN of monthly Azure allowance.
As a result, everything had shut down!

What does the Cloud Adoption

Framework say?
In the “Operate” phase of the Microsoft Cloud Adoption
Framework there is a section on “Govern”. The Govern section
focuses on the “Five Disciplines of Cloud Governance:” which are:
 Cost Management
100

..MS2019..
 Procurement and financial governance

 Security Baseline
 Resource Consistency
 Identity Baseline
 Deployment Acceleration
The govern section also contains a guide to “Improve the Cost
Management discipline” and some tooling to aid in the creation of
a governance baseline for your business. The Ready section also
includes guidance on how to deliver this.
For more information on cost management within the Cloud Adoption
Framework for Azure, visit: http://aka.ms/adopt/gov/cost
and
http://aka.ms/adopt/ready/managecost

Operational Cost Governance

Non-production workloads illustrate how costs can be managed as
part of your operations ITOM layer. These workloads rarely need
to be on 24/7. They can often be off for long periods. To illustrate
this concept, let’s suppose the release cycle for a vendor-supported
application is quarterly. You may need two weeks for testing prior
to release to production. But for the rest of the time the test region
for the service is idle. If this service is switched off during idle
periods, you can make savings on compute costs. Over time these
savings will mount up.
Workloads having variable, burstable usage patterns are another
case. If the patterns are known and predictable, automated tasks
can be carried out on the estate to cope with them. Nodes and
clusters can be scaled both vertically and horizontally.
In retail a burst can be an order of magnitude above the normal. A
stamp type model could be the solution here. This incorporates a
full functional cross section of an application including
presentation, application and database. In this way, even very large-
scale fluctuations can be handled. You no longer need to scale your
environment for peak demand. In this case it may just tick over at
ten percent capacity for ten months of the year.
If scale is less predictable and fluctuates over a day, auto scaling can
be considered. Auto scaling works very well in a Cloud-native

101

..MS2019..
Procurement and financial governance

architecture. This is where each component is stateless and is

designed to scale horizontally.
Another case is that of a microservices architecture. When each
microservice is designed in a Cloud-native way you can scale up
and down according to the load on the service. This is done with
set parameters within the Cloud fabric.
For heritage applications you can adopt a similar approach. But it
will be slightly less elegant owing to the inherent lack of flexibility
of the application architecture. A remote desktop service may have
ten session hosts to provide compute power to the user
community. Without the use of complex tools, you would need to
deploy all ten nodes into public Cloud. However, you could shut
eight of them down to save the compute costs. You could then
boot these servers on demand. This would be done with the auto
scaling feature. It would click in when the load of the running
nodes reached a pre-defined threshold.
There are many more use cases for cost management. They share a
common theme. Those that will be useful to your organization will
often be unique. It all depends on your scale, your geographic
reach, and so on.

Build and deploy governance

This is the second topic relating to financial governance. The cost
of services should be understood and correctly allocated at build
time. As services are built by project services teams, they begin to
incur costs.
Cost control and TCO are rarely at the front of engineering minds
in the implementation phase of a project. Why have four cores on
this server when eight is clearly better? Wouldn't it be risky to
deploy a server with only 16GB of RAM? The next one on the list
has 56GB and would be safer.
Worse still are developers who spin up multiple nodes of whatever
size feels right to test something. Then they get distracted and leave
them running for weeks unused. As organizations begin this
journey to Cloud at scale, the financial governance topic will
quickly bubble to the top of the list of priorities.

102

..MS2019..
 Procurement and financial governance

There are many ways to address this issue. Staff need flexibility to
interact with Cloud services as the requirement arises. At the same
time, someone, somewhere should be making financial decisions.
The structure of your public Cloud environment is relevant. Azure
has a combination of subscriptions, resource groups and tagging.
Careful consideration must be given to role-based access and
policies. You should also consider the use of Azure Policies to
assist in this space as you can use this feature to limit what staff can
and can’t do to the environment.
There are similarities between organizations and their use of these
features. Your final position will depend on several factors. These
include the size of the organization, the number of teams allowed
access, your geographical reach, as well as other factors. A small
team located in one place may well handle this governance at team
level. Access to the management tooling could be left largely open.
A larger team with broad geographic spread may well lock the
management console down. The same applies to one with multiple
operating companies with distributed management. It may be
necessary to limit interactions by delegation to a subscription or
resource group.
All this needs to be thought through before you put large
workloads into public Cloud.

Right Sizing and Optimization

The final topic under the heading of financial governance is
workload optimization. No matter how much modelling you do,
you will be deploying workloads into Cloud using a best guess. It
may be a very informed and experienced guess. You may well come
close much of the time. But what you are unlikely to do is watch
your workloads over time to see if you can be more efficient.
Into this space drops specialist tooling. This tooling will keep
watch on your VMs and PaaS workloads. Over time a view will be
built of how busy they are and the resources they are consuming.
This tooling will compare the resource consumption with the
SKUs available to you. It will recommend changes you can make. It
will also show the cost savings these changes will attract.
To learn more about optimizing costs based on recommendations, visit:
https://docs.microsoft.com/en-gb/azure/cost-management/tutorial-acm-opt-
recommendations
103

..MS2019..
Procurement and financial governance

The more sophisticated among these tools will also allow you to
apply personalized polices to the optimization algorithm. In this
way, recommendations you don’t want will not appear. Many of the
public Cloud vendors have a version of this capability baked into
the management portal. Microsoft purchased Cloudyn for this
purpose. This has now been integrated directly into the Azure
Portal.
There are other, third party tools in market. They may give you a
more granular estimate. They may present the data in a way that is
easier for you and your teams to understand. Some tools also have
a rich API. They can trigger run books or customized scripts if you
want to automate this activity. Many of the tools available can also
look at other procurement vehicles which you may be able to take
care of, such as reserved instances. These are advantageous in the
case of a workload that is on all the time. If you are comfortable
you know the size of the VM is correct. And provided you know it
is unlikely to change over time, you can purchase a reserved
instance. This will attract up to a 40 percent discount on the Pay-
As-You-Go pricing. We will talk further about this concept within
the capacity management section in the next chapter.

104

..MS2019..
 Service management

Chapter

10
10Service management
If everyone is moving forward together, then success takes care of itself
Henry Ford (Industrialist, 1863 – 1947)

Building a new service model

S ERVICE management is the core of your Digital Operating

Model. It is the nuts and bolts of the way Cloud will be born
and delivered in your organization. The service management
surrounding your Cloud environment will determine its
success or failure.
It’s where the
rubber meets the
road between
agility and control.
It’s your front line
between the
business and IT.
Service
management in a
traditional IT
operating model is Figure 10 - ITIL Service Map
one of the best
described and best designed set of standards. Information
Technology Infrastructure Library (ITIL) is now in its fourth
incarnation. There were major changes between v3 and v4 around
management of Cloud environments. ITIL describes every aspect
of IT service provision. It describes how to model and control it.
105

..MS2019..
Service management

More recently, Service Integration and Management (SIAM) has

emerged. This divides this service provision across multiple
providers. Different services can have a single interface to the
business.

Figure 11 - Traditional SIAM Service Model © www.itforbusiness.org

These models are widely used today. They are still described as
Future Operating Models (FOMs). Compared to what was
available before, they are brilliant. They still offer a lot of value
today. They were, however, a product of their time. And time
moves on – fast...
Many organizations now feel suffocated by this siloed approach to
IT delivery and accessibility. Standardization delivers a lowest
common denominator approach to IT. A service catalogue should
support a request for a server in the same way as a request for a
laptop. A user having a problem accessing e-mail should not be
treated differently from a user who is unable to run a CI
deployment script.
Many Cloud service lines also cut across their parent towers.
Datacenter and hosting are examples given above. How do you
deal with these new service lines? How can you present them to the
business with the best of what you do today? The business requires
more flexibility and agility to utilize the Cloud effectively. These are
some of the ideas we explore in this chapter.

A tower or a service line?

If you operate with some form of SIAM model, you’ll be familiar
with the concept of towers. A tower includes several service lines
106

..MS2019..
 Service management

and interfaces to the SIAM owner. In a Cloud world, these service

lines will often be described differently. They fail to sit neatly
within the ontology that your current model is built upon.
With a Digital Operating Model there is a relationship between the
business and the new Cloud service lines which was not there with
a traditional ITOM. How do users provision services? How do they
access them? How do they get support for them?
In a traditional ITOM, from a datacenter perspective, the server
ruled the world. Organizations have millions of server definitions
in their service catalogues. Do you want a blue one? Four CPUs or
two CPUs? How many disks? How much RAM? Once you’ve
chosen your server, you choose which environment you want it
deployed into. These include Dev, UAT, Production. There will be
pages and pages of options of configuration items (CIs) to deploy
into. They will all be managed within your organization’s ITOM.
In the new Cloud world, CIs become less important and less
relevant. This is so particularly in a Cloud-native Gartner Mode 2
world. The physical and virtual manifestation of an application at
any time of the day, week or month matters only from a billing and
monitoring perspective. It does not matter from a support or
availability perspective. Environments become ephemeral and
short-lived. They can be recreated automatically on the fly. Why
have just three environments for an application? You can have as
many as are needed by whoever is testing which part of the system?
The key to building a successful Digital Operating Model is to free
it from the confines and control of a traditional ITOM. At the
same time the security and policies that your organization requires
can be kept in place. The service line you’re building has some
similarities with a traditional tower. It still forms part of a
distributed service delivery model. But there are important
differences. It needs to exhibit some of the same capabilities, such
as SLA delivery, for example. But it can deliver this in much
improved ways – in real-time, for example.
You’ll need to understand how your tower model evolves with a
move to a Digital Operating Model. How will you expose the new
Cloud service lines? You probably won’t know the answer to that
question yet. But the answer will emerge as you expose the
underlying Cloud platforms. It will become clearer as you
understand more about how users are consuming and demanding
107

..MS2019..
Service management

support on these capabilities. The general rule is the more you

refine process the more agility you can deliver.

Service Portfolio Lead

To begin on your journey to a new Digital Operating Model it’s
useful to start to think about the notion of a service portfolio. How
can you start to structure and organize the different services that IT
might want to present up in a way that can be better understood?
How can you create stronger links between demand and value?
End User Compute doesn’t mean much to a business consumer of
an internal IT service. Teamwork does.
Consider whether you might be able to use a model that more
closely resembles this:

Figure 12 - Portfolio-based model

Grouping your services into this kind of portfolio-based model

makes it much simpler for others to consume and more closely
aligns with the delivery model from the underlying Cloud vendor.
In the authors’ view, the role of IT needs to change in this new
world. IT should not be building and operating this suite of
services itself, as it may have done within a traditional tower model.
IT should be presenting up these underlying capabilities as a set of
easy-to-consume services. IT is now a shop, not a manufacturer.
It’s important to match the portfolio of services closely that you
want to provide to the services presented by the underlying
platform. This is one of the (many) reasons why a multi-Cloud
approach presents significant additional complexity. How do you
map the items within your service portfolio to common services
exposed by the underlying Clouds? Whilst VMs may be easy, as you
get higher up the stack the differences become greater. SQL Azure
108

..MS2019..
 Service management

and Amazon RDS provide the same kind of service, but the
implementation (including performance tiers and pricing models)
vary greatly.
You will also need to consider the process you go through to
“approve” underlying Cloud services at the time you onboard them
into your portfolio of services. You will need to build governance
and compliance shims around them. You may support SQL Azure,
but only if Transparent Data Encryption is turned on. Your SQL
Azure Offer will therefore be a standard SQL Azure service, with
that setting enabled and enforced.

Provisioning
Giving users access to your Cloud(s) is central to the effective
operation of your Digital Operating Model. It comes with a fine
and delicate balance, however.
How your users get access to a Cloud environment is one of the
most contentious and hotly debated issues with Digital Operating
Models. Cloud is automated and self-service. That is one of its
defining characteristics. Nevertheless, uncontrolled access directly
to the underlying Cloud environment presents serious difficulties
for others. How are we going to control this? How are we going to
secure this? How are we going to financially account for this? In
the everlasting contention between agility and control, provisioning
is the tip of the arrow. We want users to have quick and simple
access. At the same time, we must enforce policy and governance.

Creating a Landing Zone

The first step to provision, is to build the underlying environment
you’re going to provision in to. In the past we have called these
Azure Foundations, Virtual Datacenters or Azure Scaffold. Today,
we refer to them as Landing Zones.
For details of building your Landing Zone, visit:
https://aka.ms/adopt/landingzone

Within a Landing Zone will be a core set of capabilities, networking

access control, and templates. For the purposes of this we need
only assume that we now have a core set of services available to us.
We’ll talk more about the services that would be deployed into a
Landing Zone in the next two chapters on security and tooling.
109

..MS2019..
Service management

Once we have this Landing Zone, we can create environments to

sit alongside it which we can make available to product teams. We
can then authorize that team to perform a set of actions, subject to
the governance controls baked into the Landing Zone.
This is a useful place to interact with your service catalogue.
Instead of individual configuration items in the catalogue, you can
allow users to provision an entire “Cloud Environment” connected
to a Landing Zone. During the provisioning process you collect
information from the user, such as:
 Details of the owner
 The project code
 The cost code
 Information security classification of the data
 The service wrap / operational requirements for the
environment
 Other pertinent information
This information can then be used to deploy the Cloud
environment automatically. At the same time, it can set up the
access control to the environment. Specific policies can be
deployed on to the environment, such as tagging and cost
allocation.
There are several ways to deploy these environments automatically.
Some organizations choose to treat an environment as a resource
group. They set up policy and access control accordingly. Others
choose to deploy a complete Azure subscription. There are several
factors to consider in making the choice.
Once a user has access to one of these Cloud environments, you
need to build a set of mechanisms for things to make use of them.
Different actors who will require access to these services in
different ways.

110

..MS2019..
 Service management

Figure 13 - Provisioning paths

Developers will want to consume these services from automated

pipelines. They will therefore want to be presented with a set of
curated “building blocks” which can be composed into templates
within their deployment pipelines. These pipelines will execute
many times over. You might present these to your users as NuGet
packages which they can download and embed into their
applications. You might house the templates in the same
repositories that the developers are using.
IT administrators will likely want to go to a curated set of
templates. They can use these to create artefacts or resources as a
one-off operation as and when they are required. These would
typically be made available within the Azure Portal or from
PowerShell.
End users will want access to a simple to use, intuitive web
interface to request new services, such as a Power BI license.
Whichever route “customers” of the service line take, the same
automation, the same release process and the same monitoring and
improvement should take place.

Service Catalogue
In an ITIL-based view of the world, there is only one entry point
to services. That is the service catalogue. ITIL defines each of the
services which users can consume. It gives them a shop from
where they can obtain them. Do you want a VM? Sure, select one
of the options below. Do you want a database? No problem, we
111

..MS2019..
Service management

have some available for you. The service catalogue can define
workflows, scripts, approval steps – everything a user could
possibly want. These will all have varying degrees of automation
behind the scenes. It all sounds perfect, doesn't it? Well
unfortunately, not quite!
Service catalogues can still play an interesting and important part of
your Cloud provisioning process. However, the huge volume of
items available within a platform such as Azure makes them
incredibly unwieldy. Consider for a start that there are over 200
Azure services. Then remember that there could be several dozen
variants and options for each one. For VMs alone (one of the
Azure services), there are more than 150 different ranges and sizes
to choose from. Over 150! Every week a new VM size is launched.
Every few months a new range comes online as hardware is
deployed in the datacenters. Each type of VM has specific rules as
to what disks, networks or other services it can attach or connect
to. Trying to maintain a mapping between your own service
catalogue and the underlying platform is like bailing out an ocean
with a bucket. It is theoretically possible, given enough time and
enough resource. But it is as close to impossible as you can get. It is
also a never-ending job. Suppose you managed to map the range of
services available and create deployment scripts for each one; by
the time you had finished, half of them would be out of date.
There is another disadvantage to basing your provisioning on a
service catalogue. It flies in the face of the automation that the
Cloud delivers. A service catalogue is usually on the web. A user
searches for an item and deploys it. Suppose your application has
many different items – VMs, databases, and so on. It is boring and
laborious to deploy them one by one to each environment. Some
organizations do, however, choose to make a small selection of
configuration items available via a service catalogue. That might
include a dozen VMs and a couple of database SKUs. That offers a
few easy-to-consume items so that less technically-savvy users can
get up and running. Later on, these users can be introduced to the
wonders of templates and automation.

A world built on code

Instead of providing users access to provision individual services, a
better approach is to construct a library of templates and scripts.
112

..MS2019..
 Service management

These can then serve as a useful abstraction between your users

and the underlying Cloud platform to enforce your own security
standards and policies.
Using Azure DevOps, you can create a repository of all these
assets. You can combine a set of Azure Resource Manager (ARM)
templates, Azure Policies and Azure Desired State Configurations
(DSCs).

Figure 14 - Code libraries

As an added benefit of using Azure DevOps, you can run a set of

automated tests, including the Secure DevOps Toolkit to
proactively make a set of recommendations to keep your scripts
and templates secure and in line with Microsoft’s best practice.
Users can then execute these templates directly, or they can
compose them into a pipeline of automated actions.

Incident
Incident management is as important within your Digital Operating
Model as it was on premises. Things will still go wrong. You need a
process to govern this. A move to the Cloud, however, affords far
more opportunities to automate this incident management
approach.

113

..MS2019..
Service management

Your incident management process will typically have a couple of

different variants. This depends on whether it is a response to an
incident on an infrastructure-based (Mode 1) environment or a
code-based (Mode 2) environment.
In a Mode 1 world, incident management will still take a feed from
an alerting platform. This will either invoke a run book or pass this
alert on to a technician for resolution These human resolver groups
will undertake work broadly analogous to the work that they
performed on premises. They will make use of problem
management tools to identify, triage and ultimately resolve these
incidents. Your problem management tooling will necessarily
evolve to offer more automated resolutions. Rather than describing
a set of steps which can be performed to resolve the incident, they
may reference a script or a runbook which can be executed instead.
Within your application incident model, you will be able to make
use of extensive automation. The idea is that is it easier, quicker
and less risky to re-run a deployment pipeline to build a known-
working environment of something rather than attempt to fix a
broken one. If this doesn’t work, you can roll-back to the last
release. It’s possible that there might be a bug lurking within the
newest release.
Whether it’s an incident on your infrastructure or an incident on
your applications, the amount of data and telemetry that you’re able
to pull from the underlying platform should make it much easier to
spot trends and reduce resolution time. Many organizations are
now starting to build sophisticated machine learning models. These
can predict incidents in advance. They make far more detailed
recommendations to resolver groups about suggested steps to take.
As you design your incident management approach and SLAs it’s
vital you consider compound SLAs. Your team may have an SLA
of 15 minutes to respond to and triage an incident. If the issue
turns out to be with your Cloud provider, the clock on their SLA
only starts ticking when the incident is escalated to them. They may
take another 15 minutes to respond and triage. Ensure you don’t
over-commit in terms of SLAs given the dependencies you will
have on others in this new world.

114

..MS2019..
 Service management

Change and risk

As with incident management, change and risk management need
to be re-imagined in our new operating model. Change
management is probably the most affected of the service
management processes. You must come to terms with the fact you
no longer have operational responsibility for the whole
environment. The move from build-and-run to consuming services
is transformational for change management.
The underlying service fabric for many of your services is
constantly changing. You will be able to manage some of them.
Others you will be told about in time to prepare. But some changes
just happen. Microsoft is very good at warning about upcoming
changes that they expect to affect customers. Examples are a
service reaching end of life or being deprecated. They will give you
several months’ notice and suggest an upgrade or migration path.
They will often circulate a change such as an uplift to mailbox size
in Office 365. Likewise, with a change of a licensing plan. It is the
responsibility of the in-house IT team or the service provider to
keep up to date with these changes. Changes to the underlying
platform are not exposed. They are run internally by the Cloud-
hosting provider.
You will need to look at your environment in groups of services.
There are the services you control such as your VM guest operating
systems and SQL servers. They can be managed in a traditional
way. But services you consume such as Office 365 and PaaS
services in Azure will need to be treated as third party. You will
have little or no say over when changes take place. You will,
however, be notified of these changes so you can warn staff.
Everything we have covered so far is challenging. But it is
achievable for your change process. What does this look like in a
Mode 2 world? In Mode 2 change is the whole point. You can
make thousands of micro changes a week. This is the source of
business benefit that Mode 2 workloads deliver. In such a model
you can’t possibly hope to govern these workloads through a
formal change process. Therefore, these workloads need to have
strong testing and rollback baked into the deployment process. We
cover this in more detail in the DevOps chapter. For now, we will
only say that you should educate the service organization about

115

..MS2019..
Service management

their future. They need to get a feel for how things will look as you
progress on your journey to Cloud adoption.

Capacity Management
Capacity management has a lesser, yet still important role within
your Digital Operating Model. Capacity management no longer
needs to have regard to physical equipment. We no longer need to
concern ourselves with the procurement and deployment of
physical servers and storage arrays. For all intents and purposes, we
can treat the Cloud as delivering infinite capacity. There are,
however, a couple of caveats to this approach.
This first caveat is if you are delivering a hyper-scale service within
the Cloud. If you typically need to spin up thousands of cores on a
variable basis you may hit constraint limits. Even hyperscale Clouds
only have a certain amount of excess capacity within the platform.
If you think you might fall into this category, engage with your
Cloud provider to understand how you can pay for the ability to
deliver these kinds of peak expansion.
The other interesting dimension is reserved instances. We covered
these within the financial governance chapter. Reserved instances
allow organizations to pre-commit to certain levels of usage.
Instead of purchasing a server by the minute, if you know the
workload will stay static for a year or three years, consider
committing to this. The savings will be substantial. In order to
assess these forward-looking capacity demands you need – you
guessed it – capacity management. Don’t get too hung up on
reserved instances at the beginning and certainly don’t commit to
them until your given workload has been running a month or two.
Until you can see how your application performs in the Cloud you
don’t know what size of servers you need. You don’t know
whether or not you might in fact be better off with a variable
compute model.

The new “run” business

Major changes are required to your service management
organization to support this new world of change and automation.
Barriers must be broken down. Teams need to come together and

116

..MS2019..
 Service management

collaborate more. New service façades on top of SaaS services

must be introduced. Platform services teams need to be born.
For an introduction to how to reorganize your teams, visit:
https://aka.ms/adopt/organize

Consider a traditional service such as e-mail. A simplified view of

your service management organization may look like this:

Service Desk

End User Compute Servers

Devices Software Networks Storage Compute Security

Figure 15 - Traditional support model

After moving to a service such as Office 365, your service

management organization should turn into something like this:

Service Desk

Endpoints SaaS Services

Devices Software Microsoft

Figure 16 - SaaS support model

A simple change, but one a having profound impact. There is now

a symbiotic link between your service organization and that of your
new Cloud provider. We discussed this in chapter 8. You no longer
have access to server teams to diagnose and fix issues with the
underlying infrastructure supporting your business. Your Cloud
provider now employs these people. What’s important is the
interface between you and them.
You may choose to interface directly with these providers. Or you
may choose to work via a Cloud solutions provider. Whichever way
you chose to solve this challenge, it will be regularly tested within
your new run organization. You must ensure that you have access
to the right people, as quickly as possible. Tight SLAs are required
to deliver these new services to the users.

117

..MS2019..
Service management

Support for internal line of business applications becomes even

further removed from what it is today. At present, there is a clear
separation between the development and onboarding teams on the
one hand, and the operations team on the other. There are defined
hand-over mechanisms.

Figure 17 - Traditional application support model © Servian

But this disjointed approach to application delivery is slow. There

can be a delay of four to six weeks in getting an application from
the project team to the operations team. When issues do arise,
there can be a them and us conflict between the two teams. Once
the project team has delivered, they often refuse any responsibility
for issues in production. This increases the complexity of the
handover. The operations team try to foresee every problem that
might arise later. They want to satisfy themselves that the
application is completely ready. In our new run world these teams
come together. They become two sides of the same coin, working
collaboratively together:

118

..MS2019..
 Service management

Figure 18 - Next generation applications support © Servian

Key to this new support model is the platform services team. This
team is the special sauce. They are the ninjas, the automation
specialists. This team mediates the work of the applications team
and the work of the operations team. They support both internal
applications teams and third-party ones. They work between the
two worlds.

Structuring your new teams

There are several approaches to structuring your new teams. Within
the Cloud Adoption Framework, Microsoft suggests at least two
teams. A Cloud Adoption team and a Cloud Governance team.

Figure 19 - Cloud adoption and governance teams

For more information, visit:

https://aka.ms/adopt/organize/establishteams

119

..MS2019..
Service management

These two teams will naturally compete. That is OK. The Adoption
team will be focused on identifying new opportunities to leverage
Cloud technologies. The Governance team will be focused on
keeping them safe and secure.
For each of these teams, you can choose the center of excellence
model, or the community of practice approach. Both have had
some measure of success in organizations today. There are many
aspects to the choice of model for your organization. Issues to
consider include your funding models and the overall roles and
responsibilities in your organization. There are also the DevOps
principles that will mandate the reduction or removal of silos.
The community of practice approach is suitable for organizations
that are product and green fields based. The development teams
will complete the building of the new capability. They will ensure
that it’s made ready for production. They will also ensure that it is
clear what the developers will look after and what operations will
be responsible for. They will then return to their product teams.
A center of excellence style platform services team is suitable in
other situations. For example, in a large enterprise, development or
product teams could appoint members on a more permanent basis.
That would allow them to lay out their requirements and support
the ongoing provision of services.
For an example RACI matrix for these teams, visit:
https://aka.ms/adopt/organize/raci

Over time, there will be a requirement to align these teams to your

existing central IT teams. This is not without risk. The central IT
team must come to the party with a growth mindset. The worst
situation is where these teams gravitate back to traditional
approaches. They might fall back on traditional ways of doing
things to “align” with the greater IT service management universe.
You must not let them. Central IT needs to mimic the new model,
not vice versa.

Starting from the top

All these fundamental changes to your service management
organization must start from the top. They must start from the
beginning too. As project/innovation and operations teams come
closer together, so do their reporting lines. Digital initiatives within
120

..MS2019..
 Service management

organizations often drive a lot of this change. But they must

operate closely with operations. Have you ever heard about a digital
team who decreed PaaS-first, while the operations team didn’t
officially support PaaS in production? You come across it time and
again. There are two poles pulling in opposite directions – two
organizations without a common strategy or goal.
You want to get your service management organization firing on all
cylinders and adequately supporting the business. To do so you
must dissolve the fiefdoms. Everyone needs to be on the same
page within the one organization.
For more information on dissolving the fiefdoms, visit:
https://aka.ms/adopt/organize/antipatterns

There are different approaches to achieve this. The simplest is to

make sure everyone has the same reporting lines. Several roles can
perform this function. It might be a Chief Information Officer
(CIO), a Chief Technology Officer (CTO), or a Head of Digital
(HOD). It doesn’t so much matter the title given to this person.
What matters is that this person heads a joint organization. They
must ensure that the different parts of the business play nicely
together. This is essential if you want to become truly digital.

121

..MS2019..
Service management

122

..MS2019..
 Access control, security and policy

Chapter

11
11Access control, security and
policy
The need for security often kills the quest for innovation
Haresh Sippy (Industrialist, 1946 –)

Rethinking your attitude to risk

S ECURITY, security, security! Never has one word struck so

much fear in the minds of CIOs, CEOs and boards of
directors. It’s a trump card that can be played in an almost
arbitrary way in any situation, but it is often played to achieve
the wrong result.

We can't  Why not? SECURITY

do this
We can't  Why not? SECURITY
do that
In any conflict between security and innovation, security wins.
Security is, of course, incredibly important. You only need to watch
the news to hear about the latest data breach, the latest
ransomware, the latest exploit discovered in something you use
every day. In the words of James Rozoff: “[Although] security is an
illusion … it is a pleasant one.” Security is ultimately about risk
123

..MS2019..
Access control, security and policy

management. What level of risk are you prepared to accept as an

organization? What odds are you prepared to place a bet on? A one
in a million chance of something happening, or a one in several
billion? What damage might arise from the theoretical threat we are
protecting against?
For many years, security was the main reason organizations gave
for their failure to adopt public Cloud services. We can’t use service
xyz because [insert some reason related to security]. Over recent
years, most of this fear, uncertainty and doubt (FUD) has melted
away. People now admit that the security delivered by most public
Cloud providers is actually superior to the security that can be
delivered on-premises. But many still don’t admit it, and still
believe the Cloud-is-not-Secure mantra. To those readers, you may
as well stop reading here. We wish you good luck. The rest of this
chapter is addressed to those who are willing to trust that the
public Cloud can be configured to deliver security – security that is
at least as good, if not better than what is available on premises.
You will now want to learn how to take advantage of the security
that is on offer from the public Cloud. In this chapter we will
explore the things you need to take into consideration and the
questions you should ask. In that way you should be able to deliver
something at least as secure as you can deliver today, and more
likely, more secure.

Access Control
Access control is authentication plus authorization. That is to say,
identity is the most important ingredient in the recipe for what are
you allowed to do.
For a guide on where to start with your identity baseline, visit:
https://aka.ms/adopt/gov/identity

Identity as your chain of command

A central part of any Digital Operating Model is the definition and
implementation of a Cloud identity provider (IdP). Your operating
model depends everywhere and in every scenario on the system
that identifies who you are and/or what group of users you belong
to.

124

..MS2019..
 Access control, security and policy

In the on-premises world, Active Directory is the source of this

identity almost exclusively. Every user and every service have an
Active Directory credential. In the Cloud an equivalent of Active
Directory is needed. But instead of authenticating and authorizing
access to devices and legacy applications, a Cloud identity controls
access to Cloud platforms and services. A Cloud IdP is a single
sign-on (SSO) platform on steroids.
There are several Cloud IdPs on the market today. They include
Okta, Auth0, Centrify, and Azure Active Directory. Each has its
pros and cons and an associated commercial model. Most deliver
some form of Active Directory federation. The first thing to do is
undertake an evaluation of these IdP platforms. Look at what they
support and ask the smart questions. Choose the one that will work
for you. Then go ahead and deploy it as quickly as you can.
Deploying a Cloud IdP can be straightforward, or it can be
complicated. If you’re linking it to your existing on-premises Active
Directory, additional work may be required. You may need to tidy
up your current environment since you might not have followed
best practice. Your environment might have evolved over time.
You might have acquired or divested organizations which form
part of your Active Directory forest but are not good neighbors.
Although sometimes daunting, this work must be done. Your
identity model forms the core of your Digital Operating Model.
Take the time to do it right. The good news for Office 365 users is
that you already have a Cloud IdP. You may not realize it, but the
page where you log on to Office 365 is in fact Azure Active
Directory.

Multi-factor authentication
Once you have established your IdP, you need to build the controls
which surround it. Multi-factor authentication (MFA) used to be
the preserve of the biggest organizations and those most paranoid
about security. With the Cloud, this capability has been opened to
everyone. The need for expensive hardware access tokens or
proprietary solutions has gone. It’s typically built in, out of the box.
You now have the choice of rolling it out to all users or to a subset.
Another option is to base it on a set of criteria. Such criteria might
include whether the user is at home or at work, or whether

125

..MS2019..
Access control, security and policy

suspicious activity has been detected. Agree your rules and

incorporate these into your Digital Operating Model.

Single sign-on
MFA was the exception rather than the rule. The same is true of
single sign-on. It used to be a nice-to-have option. It was
something you did if it wasn’t excruciatingly difficult or expensive.
Within a Digital Operating Model, it’s a virtual necessity.
We discussed earlier the requirements you place on service
providers as part of a procurement exercise. SSO support must be
a major factor in your decision. SSO standards such as OAuth have
proliferated. And so, finding a service provider who can participate
within your identity realm is relatively straightforward.
Many Cloud IdPs have support for hundreds if not thousands of
pre-built integrations. At the time of writing, Azure Active
Directory supports more than 2,500 SaaS-based applications.

Joiners, movers and leavers

JML is an important process to introduce into your Digital
Operating Model. A JML process describes the steps to be taken
when an employee starts at your organization, changes role or
departs. Giving people access to systems, changing their access, or
withdrawing access all have serious security implications.
The work to be done depends on the kinds of Cloud services and
Cloud IdP you are using. It also depends on the setup of your
single sign-on access to each system. Typically, single sign-on is
based on identity federation and a token-based chain of control. A
target system uses your main credentials to provide access. For
these services, your JML process will be fairly streamlined. You will
create or disable a user account. This will be done either within
your on-premises Active Directory, or directly within your Cloud
IdP. Users can be moved into, out of, or between security groups.
This is not very different from the way you do it on premises
today.
In some scenarios, relying services have their own instance of a
user’s identity. This will be piggybacked onto the main credential.
Enabling or disabling a user’s primary credential doesn’t necessarily
enable or disable the secondary account. Moving a user between
126

..MS2019..
 Access control, security and policy

security groups might not modify their access within these systems.
For such scenarios, you will probably need to increase the
complexity and tooling that your JML process leverages. Suitable
technologies include Microsoft Identity Manager (MIM). There are
others. Azure Active Directory Premium (AADP) also has prebuilt
functionality to achieve this for Cloud services such as Workday.

Best practice for your cloud IdP

At this point you have your Cloud IdP enabled and deployed and
your users synced with it. Now you can enable federated access to
your Cloud services, such as Office 365 and Azure. The best
practices you have employed on-premises for years apply just as
well in the Cloud. One such practice is to always use security
groups rather than individual users when giving access to
operations within Cloud environments. Leveraging inheritance and
hierarchies is an example. Start with a broader definition of users,
then narrow it down. Definitions of users and the kinds of
operations you want to allow them to undertake become more and
more specific. These concepts are no different in the Cloud and
should be the same in your Digital Operating Model.
For more detailed guidance on the set of Azure tools available for Identity,
visit: https://aka.ms/adopt/gov/identity/toolchain

IdP and your cloud platform

For SaaS applications, your identity model will be relatively
straightforward. It will broadly mirror the way users are provided
with access to on-premises applications today. For your Cloud
platform (IaaS/PaaS) the complexity increases somewhat. Your
IaaS and PaaS layers will contain the actual nuts and bolts of your
underlying applications. This includes the servers and services
which compose them. You will typically have multiple
environments for each of your applications, including test/dev,
UAT and production. Differing user access rights will apply to each
environment and to sets of services.

127

..MS2019..
Access control, security and policy

Figure 20 - Azure security model

The concept of an environment may also change. With traditional

infrastructure, you have pre-provisioned environments. They are
set up in advance and into them servers and security groups are
deployed. You will still have these kinds of environments in the
Cloud. But you may also introduce ephemeral environments. An
ephemeral environment is one which is created and destroyed on
the fly, often at a high rate. The environment will be declarative.
That is, it is self-describing and encapsulated. It will be based on a
set of templates and scripts which can be automatically deployed.
An application won’t be deployed into the given environment.
Instead a new test and dev, UAT, or production environment will
be created. This will be done on the fly, specifically for this
application. It’s a very different concept from what you have been
used to.
What makes up an environment may also be different in this new
Cloud world. Traditionally, an environment might include a large
set of assets, including servers, networks, VLANs, switches, and so
on. In the new Cloud world, an environment might just be a
defined group of services, such as a resource group.

128

..MS2019..
 Access control, security and policy

Figure 21 - Azure role-based access hierarchy

However, your environments are being deployed, you need to

make sure you have sufficient identity controls to secure and
manage them. Typically, this will comprise owners and a set of
differing roles beneath them.
For more detailed guidance on RBAC controls, visit:
https://aka.ms/adopt/ready/manageaccess

129

..MS2019..
Access control, security and policy

Security
The security of your Cloud environment is all-encompassing. This
is a broad topic. Your approach to security will involve several
individuals and teams from across your organization.
Normally security covers physical, infrastructure, networks and
VM/apps. After moving to the Cloud, you no longer need to be
concerned about the physical and infrastructure layers. Instead you
can focus your efforts on the network and VM/app layer. Your
security design will not necessarily form part of your Digital
Operating Model. But the principles and vendor solutions (native
vs third-party) that you will enforce do need to be agreed and
planned. They will have far-reaching implications for things such as
support, performance and availability.

Figure 22 - Trust and control model

Perimeter security
You need to protect your perimeter first. That is no different from
a traditional on-premises environment. Azure contains several
capabilities built natively into the platform to help supply perimeter
security. Among these are Azure’s DDoS Protection and Mitigation
service and its Firewall service.
There are also third-party security products on the Azure
marketplace. Among these are Palo Alto, Barracuda, and
Checkpoint. There are others. These third-party appliances are
useful when designing a hybrid perimeter security model. Especially
so if you have physical or virtual appliances from these vendors
deployed on-premises today. In this model, Azure becomes just
130

..MS2019..
 Access control, security and policy

another datacenter with ingress and egress points needing to be

protected.
Customers who have deployed ExpressRoute have additional
options and requirements for their Azure environments. An
ExpressRoute circuit is a private link between the Microsoft
datacenter(s) and your existing offices and datacenters. Some
customers chose to route all their traffic into or out of Azure and
Office 365 via this connection. In this way, all traffic can be
inspected on-premises via an appliance. The exceptions are certain
whitelisted internet services such as Windows Update.

Figure 23 - Azure ExpressRoute conceptual diagram

131

..MS2019..
Access control, security and policy

Leveraging ExpressRoute and forced tunneling can have significant

impact, reducing your bandwidth and network capacity. It is
something that should be carefully considered.
You will also need a web application firewall (WAF) if you are
exposing web applications to the internet directly from inside
Azure. A WAF delivers more than a standard firewall. It does more
than answer the question: Can I connect on this port? A WAF
understands the detail of web traffic. It inspects traffic for
inappropriate behaviors and attacks by malicious third parties.
WAFs protect against common attack vectors such as SQL
injection attacks.
Dedicated web application firewalls used to be expensive to
procure and to configure. In Azure, they are even simpler than
deploying a VM. Azure has its own WAF capability built in called
Azure Front Door. Alternatively, there are third parties offering
solutions out of the box. Both allow extensive configuration and
control of web traffic entering your environment.

Internal security
Now that your perimeter security is in place, you need to ensure
that you enforce appropriate separation of network segments
internally. This is something you may have done using VLANs or
by enforcing security between different subnets. In the Cloud, this
is still what you do. Only now you have a different set of tools and
services. Instead of VLANs, we have virtual networks in Azure.
Each virtual network is assigned a pool of IP addresses and a set of
routes and connections through to other virtual networks. In Azure
you set network security groups. This is like setting ACLs on a
firewall to control traffic between network segments, These NSGs
can either be applied to VNETs or to resources which attach to a
virtual network.
Some organizations choose to deploy third-party appliances to
control the flow of traffic between virtual networks. That is in the
same way as they control ingress and egress from a Cloud
environment. Others choose to use native capabilities.

132

..MS2019..
 Access control, security and policy

VMs and apps

Security at the VM and application layer is complex. It is a separate
topic that requires to be considered in depth. Each application will
present a bespoke requirement for connectivity. The approach you
take to secure each one will be different.
The main issue for your Digital Operating Model is not the
specifics. It is to define your overall approach to securing your
VMs and applications. Many organizations moving to the Cloud
take the opportunity to apply additional controls. They might not
have been able or willing to apply such controls before. These
might include:
 Host-based security / firewalls
 Enforcing IPSEC-based connections to VMs
 Deploying advanced antivirus and host-based intrusion
detection
It is unnecessary for your Digital Operating Model to get into
details at a VM by VM or app by app level. But it does need to
include some high-level objectives and controls. These are the
controls you would like to be introduced. This leads neatly on to
the next chapter about monitoring, management and automation.

Azure Security Center

Now you’ve followed best practice and baked security into every
part of your Azure environment. How do you police this and check
that your environment is configured correctly? The answer is the
Azure Security Center (ASC). Despite being a Cloud service, ASC
can provide visibility across your entire hybrid estate. It can help
manage policy compliance. It can also perform security assessments
and make proactive recommendations on issues to be addressed
and resolved. Find out how you can leverage these capabilities as
part of your Digital Operating Model.

Microsoft Threat Protection

We have been looking at the work to secure your new Azure
datacenter. There are other tools and capabilities to help secure
your end-to-end Microsoft Cloud environment. Microsoft Threat
Protection (MTP) is one of the newer ones. MTP provides a view
133

..MS2019..
Access control, security and policy

of an organization’s overall threat landscape. Administrators can

easily spot new threats and attacks. MTP pulls data from numerous
services and combines them for viewing on a centralized dashboard
– such services as Office 365 Threat Intelligence, Azure Active
Directory Identity Protection, and Windows Advanced Threat
Protection.
The most interesting aspect of Microsoft Threat Protection is the
sophisticated artificial intelligence baked into it. New and unseen
threats can be spotted and stopped in real time. Embedding MTP
in your Digital Operating Model will provide additional security
and assurance for your Microsoft Cloud environment.
Microsoft Threat Protection also hooks into Microsoft’s brand new
Security Incident and Event Management (SIEM) product,
Sentinel. Sentinel is a class-leading Cloud-native SIEM which
provides limitless storage and analytics for security signals from
across your organization.

Policy
One of the most fundamental shifts within an effective Digital
Operating Model is to move from a process-based model of
security and compliance to a policy-based one. What do we mean
by this?
Policy and compliance are one of the three bedrocks behind Govern within the
Cloud Adoption Framework for Azure. For more information, visit:
https://aka.ms/adopt/gov/corporatepolicy

Today, most organizations’ approach to security and compliance is

to leverage processes. Detailed processes will be drawn up for
every aspect of IT security and compliance. These processes will
describe everything that has to happen, step by step for an
application or service to be approved and deployed into
production. This process will typically involve dozens of signoffs,
design committees, accreditations and more. This process-based
approach is largely successful and ensures that there are many eyes
on things as they go through the process. Multiple points to catch
bad designs or people trying to do things which would impact on
security or compliance. There are, unfortunately, three major
challenges with this approach:

134

..MS2019..
 Access control, security and policy

Firstly, it is highly reliant on people. As we all know, people make

mistakes.
Secondly, especially in terms of the successful adoption of a Digital
Operating Model, it is slow, laborious and complex. Trying to build
a process to govern the use and deployment of any of the hundreds
of services available within Azure verges on the impossible.
Thirdly, and most critically, it’s very difficult to demonstrate
compliance with a process-based approach to security and
compliance. Sure, we can say we have a process. We can say that of
course everyone follows the said process. But, as we know, those
two statements alone are not always enough to placate accreditors.
What we need is a deterministic methodology to demonstrate
compliance. Consequently, we suggest a different route. Enter
policy.
With a policy-based approach to security and compliance we switch
the model on its head. What if, instead of describing a process
which enforces checks against controls, we could describe the
controls themselves. Rather than having a technical design review
to check that there were sufficient separation of controls, we had a
policy which could describe it. Rather than have a committee check
that all aspects of the application are deployed into a certain
sovereign geography, let us have a policy that enforced it. That is
the core of a policy-based approach.
With this policy-based approach we have several significant
advantages:
Firstly, we can start to achieve the balance between agility and
control which is the centerpiece of our Digital Operating Model.
With a policy-based approach the control can be expressed as
policy, not a series of steps in a process. An environment can be
evaluated against a declarative policy in seconds. A process could
take days or weeks. Once we are comfortable that the policy is
enforcing the control, we can start to scale back the process. We
can even, once we’re really comfortable, do away with much or all
of the process. The policy serves as a set of guardrails to keep
everyone safe. We no longer need to keep people safe with process.
The second and even more powerful aspect of this policy-based
approach is its ability to demonstrate, deterministically, whether
your environment is compliant or not. If our control is expressed
135

..MS2019..
Access control, security and policy

as a series of policies, we can say definitively that we are either in,

or out, of compliance. We don’t need to even describe a process
anymore. The fact that the policy is there and that everything is
compliant demonstrates that we can rest safe in the knowledge that
all is in order.

Creating your policies

The native tool to achieve this is Azure Policy. Policy is the
mechanism you use to control and enforce the provisioning and
ongoing security and management of resources within your
environment. Policies allow you to define what good looks like.
They define the settings, configuration, placement and many other
rules that apply to your Cloud(s). Some of these policies will be
simple to conceive.
 I want all my data stored within the European Union – or
maybe specifically in England.
 I want all data to be encrypted at rest.
 I must be able to allocate usage charges to
departments/cost codes.
Some policies will be more complex. If this and this and this are
true, I want you to do this, so long as something else is not true, in
which case I want you to do that.
Policies should, wherever possible, align with security
requirements. They should also, wherever possible, match external
security standards, such as PCI DSS. In this way, you can
demonstrate security compliance through policy compliance.
Azure has wide support for policies through Azure Policy. Azure
Policy allows you to define complex sets of rules which can be
enforced at both provisioning time and ongoing. They can require
that, for instance, any VM being placed into any given environment
should have certain agents deployed and configured. They can
require that transparent data encryption is enabled for any SQL
Azure instance.
Azure also has capabilities provided by Desired State Config
(DSC). This allows detailed configuration patterns to be attached to
Azure services. The patterns apply on deployment. They also
maintain compliance with the configuration over time. There are
also popular open source solutions such as Chef, Puppet and
136

..MS2019..
 Access control, security and policy

Ansible. They can be made available within your Azure

environment. They are useful for organizations that use a lot of
open source platforms. Also check out the newly released Azure
Deployment Manager.
For details of building policy into your Landing Zone, visit:
https://aka.ms/adopt/ready/gov

Third-party products also exist in this space which can span

multiple Clouds. These products allow you to set policies which
can be applied to environments after the fact. They support several
external standards. The mappings are already done for the features
and capabilities on the underlying platforms. You might already
have an environment where you hadn’t deployed Azure Policy or
DSC. But you now want to bring it into compliance. You might
not want rigid policies. You might want to allow developers greater
freedom but have guardrails in place to keep them on the straight
and narrow. You might run other products as well as inbuilt
capabilities. This will allow you to enforce compliance and then
independently demonstrate it.
Whatever approach you take to policy, make it your central
platform. Make it the single point of authority. Keep it as fully
managed source code within your software development platform.
Protect it. Nurture it. It’s now the front line between you and the
bad guys and girls. You need it, and you need it to work. It’s quite a
significant change in mindset, but this one simple paradigm shift
has profound impact and is one of the central themes of your
Digital Operating Model.

137

..MS2019..
Access control, security and policy

138

..MS2019..
 Monitoring, management and automation

Chapter

12
12Monitoring, management and
automation
The first rule of any technology used in a business is that automation applied to
an efficient operation will magnify the efficiency. The second is that automation
applied to an inefficient operation will magnify the inefficiency
Bill Gates (Founder Microsoft Corporation, 1955 – )

F OR those that have skipped ahead, one of the most

powerful messages we have attempted to relay in the book
so far is the need for automation in the Public Cloud. The
authors are often asked to define what Cloud really means.
The answer is typically something along the lines of the application
of automation and software to tasks which were traditionally
manual. That and very big datacenters! Connecting two network
segments together used to involve a physical piece of wire. It’s now
a software command.
A Public Cloud fabric allows you to instruct the fabric
programmatically to perform a task via an API call. This can be
done from anywhere at any time through the magic of a software
defined environment. An organization with a Digital Operating
Model becomes a software defined business.
It is this critical thing which brings about the efficiency of
delivering services from the public Cloud. But it is typically the
operations function that benefits the most from automation.
It should be the aim of all businesses to run as much of their IT
operations as possible using tools rather than people. In so doing,

139

..MS2019..
Monitoring, management and automation

staff members will be freed to add value higher up the business

value chain.
We have referred several times to the Gartner concepts of Mode 1
and Mode 2. With Public Cloud there is the option to meet
somewhere in the middle with what we might call Mode 1.5. In this
mode you can automate the management of Mode 1 workloads in
something approaching a pseudo-Mode 2 manner. To put it simply,
you don’t have to transform your workloads to transform your
service management function.
Clearly, the more a workload is native to the Cloud, the easier it is
to apply sophisticated automation. In the DevOps chapter we’ll
talk about how to achieve this for software development. For now,
we will look at how this applies to Mode 1 workloads and how you
can begin to move towards a Cloud-native Operating Model.
Take the more traditional workloads such as COTS applications
installed and running on VMs in public Cloud. These can also
benefit from the type of automated processes which Cloud can
provide. To illustrate this, let’s take as an example a simple three
tier application. Our fictitious application has a presentation layer
on a web service running on server A. It also has an application
layer with some functional binaries that do “a thing” on server B.
Finally, there is a database on server C.
Using Cloud-based tools you can monitor the health of the guest
operating system. You can also monitor the primary workload
running on each server’s guest operating system. In addition to that
you can get telemetry from the Azure fabric for the services you are
consuming to run these VMs. This telemetry includes the compute
object, the storage account, the end point, the network card and so
on. This information from the Azure fabric can be blended with
the log information coming from the guest operating system. You
now have a good understanding of the state of the component
parts. Microsoft provides tools such as Log Analytics and Service
Map. These components can be stitched together, providing a
single application level view of health.
Now, we can imagine a situation where a component is in trouble
or the guest operating system is throwing warnings out. Using Log
Analytics you can create alerts that query the log data. In turn the
alerts trigger automated run books in Azure. These can execute

140

..MS2019..
 Monitoring, management and automation

commands on the fabric and the guest operating system as

required.

Figure 24 - Monitoring map

With these simple tools, we can now script almost any

configuration change on both the Public Cloud fabric and the guest
operating systems. This opens the doors to a whole new world of
support options.

Monitor / Signal / Telemetry

Key to all this happening is a common and consistent toolchain.
Organizations have over the years accumulated a vast array of tools
from different providers. Each delivers a set of capabilities to
monitor, collect telemetry from and manage an overall
environment. These include storage arrays, network performance,
and a dozen others – disparate and disconnected.
Often these tools were considered best of breed by various
analysts. Quadrants, waves – each operated very effectively with
the set of tasks it was asked to perform. Time goes by though, and
the tide is shifting.
Now a new set of tools and services present themselves. Tools and
services which you might describe as Cloud-native or alternatively
141

..MS2019..
Monitoring, management and automation

platform-native. They are tools and services which are part of the
underlying Cloud itself.
Microsoft has given particular focus to this area and has worked
hard to build a set of services to replicate its traditional on-premises
software. Many organizations today leverage the Microsoft
management suite, System Center. The new set of capabilities
which are built into the Microsoft Cloud build upon these
foundations. Yet they have been completely rewritten for the
Cloud.
What we need within our new Digital Operating Model is a holistic
approach to monitoring and alerting. A centralized approach. One
that leverages that common and consistent, but now platform-
native set of tools.
In the immediate term our focus will stay on VMs, as these are
likely to be the first workloads which you onboard into the Cloud.
Azure Desired State Configuration allows you to define a set of
configurations and tools to deploy onto machines as they are
migrated. Deploying Azure Monitor and its underlying Log
Analytics provides an extensive set of capabilities which you can
leverage out of the box. There is full, native integration into Azure
Security Center and Azure Sentinel.
This simple set of tools provides an end-to-end, enterprise grade
setup, for a fraction of what your equivalent on premises toolchain
costs today. Whilst there will invariably not be 100 percent feature
parity to what you might have today, the gap is closing rapidly. If
you haven’t got something approaching this today, you now do.
For very little extra investment.
For further reading on monitoring, visit:
https://aka.ms/adopt/ready/monitor

Alerting and triggers

Once you’ve agreed on what your toolchain will be, you need to
build an alert management workflow. An alert is raised when a
service breaches a defined threshold. The monitoring tool is then
told to act in some way. These alerts can be simple metric-based
counters, like those for CPU, Memory and Disk usage. Or they
may be more complex such as a service fault.

142

..MS2019..
 Monitoring, management and automation

It is even possible to get very sophisticated and map a higher-level

service like an application to its component parts. This will reveal
the impact of a given component on the overall health of the
system.
These alerts then need wiring into your service management tool.
You’ll need to understand what these alerts mean and write logic to
handle them. Some may be background noise, some indicative of
an actual issue where intervention is called for. In the same ways
that we do on premises today, we need to correlate and understand
the pattern of these alerts.
Initially, you will almost certainly undertake much of this work
manually. Alerts will be routed to human resolver groups for
investigation. As you and your tooling get more sophisticated you
can begin to apply additional levels of intelligence and start to
introduce automated remediation.
One of the significant advantages of leveraging Cloud-based
tooling lies in the addition of services such as best practice
analyzers. They can make proactive recommendations using the
event information and alerts. Many can be automatically
remediated, straight from the console.

Modern Management and Compliance

Monitoring and management need to extend from the Cloud to the
edge. Devices accessing your central Cloud environment need to be
governed, monitored and controlled to maintain your overall
compliance position. With Windows 10, Microsoft can now apply
the same native management experience to the edge.
This means an end of gold images. An end of “corporate approved
devices”. Now anyone can buy any Windows 10 machine that
meets a certain specification and enroll it as part of the out of the
box (OOB) experience. Users can simply sign in with a set of
corporate Cloud credentials and the device builds itself to the
corporate standard. With Windows Autopilot companies can also
pre-enroll devices to force them to join Azure AD and Intune
automatically.
With Cloud-enrolled devices, Windows now becomes “Windows as
a Service”. With this upgrade motion, Microsoft now assumes
some of the responsibility for the build, deployment and service
143

..MS2019..
Monitoring, management and automation

motion of endpoints. Rather than traditional three-year upgrade

cycles, devices now stay up-to-date with the latest version of
Windows.
Once in Intune, other policies and applications can be deployed
and the Servicing Channel chosen. Intune is an always-on, internet-
facing service. And so, every time a device connects to the internet
it will check in with Intune and policy or updates will be deployed.

Scripting
Some of us have been in IT for a couple of decades. For us in the
Unix/Linux team, scripting was a way of life. We lived in our
BASH shell and crafted scripts. We used arcane commands like
awk, grep, and sed to do all manner of things.
Microsoft brought in a Windows-based scripting language with the
introduction of PowerShell v1.0 in 2006. This replicated abilities
previously only enjoyed by Unix sysadmins. Now Windows-based
sysadmins could automate repetitive tasks. They could assess or
alter the configuration of servers and workstations which reduced
the need for tools or compiled code. This move to scripting
therefore started long before the Cloud became mainstream.
Microsoft introduced the concept of a “core” server back in
Windows Server 2008. No Windows Explorer shell was installed.
All tasks were carried out using the Command Line Interface
(CLI). Since then, all server operating systems have had this option.
PowerShell went through several iterations before the decision was
made in 2016 to take it multi-platform and open source. The Nano
server option with 2016 and 2019, for example, can only be
administered locally though PowerShell
When Microsoft developed the interfaces for both Azure and
Office 365, it was no surprise therefore that they built them on
PowerShell. The IT community now had a common language to
administer everything. That includes a Windows desktop client, a
server guest operating system, back office services like SQL server
and Exchange. They could also administer the Cloud fabric these
services run on, if they are being consumed from Azure and Office
365.
Microsoft also adopted JSON to transmit data describing types and
attributes of services in Azure. PowerShell enables you to assess
144

..MS2019..
 Monitoring, management and automation

and change the configuration of items across the whole Cloud

service stack. Add to that a descriptor file using the JSON standard
and you can pretty much do anything you like. You no longer need
to sit at a graphical interface. This approach is becoming the new
normal.
The more you advance in the Cloud maturity model and the closer
to Mode 2 you get, the more scripting becomes the only way to do
work. If you and your team don’t script today, it is vital you learn
this skill. The best way to operate your Cloud is to apply an
“everything-as-code” mindset. Development and infrastructure are
converging. This becomes more obvious as you progress along the
Cloud maturity model. Sharing code repositories from the
developers with the infrastructure teams makes a lot of sense. Build
a central repository, Azure DevOps is free for up to five users or
you could use Git. Next, get team members to collaborate together
to build out this repository. Your first scripts don’t need to be
fancy, but they should save you work.
For example, suppose you regularly refresh a test region with data
from the production system before testing a release Script it.
Having done this a few times, you may take the next step. You may
realize you don’t even need the test system most of the time. You
could consider using a script to automate the provisioning of this
test region. You have now saved both time and money. As you get
better at scripting you draw closer to the next generation of
automation and management: DevOps.

145

..MS2019..
Monitoring, management and automation

146

..MS2019..
 DevOps and application development

Chapter

13
13DevOps and application
development
Every company needs to be a software company
Satya Nadella (Microsoft CEO, 1967 –)

W E dissected in detail in Part 1 the wave of digital

disruption that is sweeping the world at present. We
looked at the opportunities and threats that this presents.
Technology has phenomenal power to disintermediate
and radically transform different industries. We introduced the
concept that every company needs to be a software company. But
what does this mean? How does a company become a software
company? How do you place technology at the core of everything
you do? Of course, there are a hundred answers to this question. In
this chapter we tackle perhaps the most complex of them. How do
you start to build a software development function within your
business and within your overall Digital Operating Model? How do
you put processes and governance around what is fundamentally a
fast moving, fast changing and highly creative discipline?
For further reading on innovation and becoming a software business, visit:
https://aka.ms/adopt/innovate

What is DevOps?
This perhaps is one of the most difficult definitions within your
Digital Operating Model. There are entire books, several times the
size of this one, which inspect, dissect and attempt to define this
nebulous word. Is it a set of tooling? Is it a process? Is it a
147

..MS2019..
DevOps and application development

redefinition of roles and responsibilities? Is it a change in corporate

culture? The answer is, all of the above. It represents a paradigm
shift in thinking about software engineering and operational
management. A very simple definition is:
The way to empower developers to produce high quality software,
which solves real business challenges, and which can be
automatically tested and deployed into a secure environment where
end users can access it
There are, of course, several orders of magnitude of complexity
surrounding part of that definition, but as a concept it is simple.
Developers write code. We need to test that code and we then need
to deploy that code. Easy, right? Not quite so easy, no!

The great war

If DevOps is what we want, what are we replacing? Grab your
sandals, it’s time to go tree hugging.
If we look at the typical software development lifecycle today, one
of the biggest challenges is that there are multiple disciplines at
play. Developers write code. They understand code and object-
oriented programming. They understand abstraction and
parallelization, and the vast array of tools and techniques available
to deliver the magic of software. Operations folk deploy code.
They understand servers and switches and security. They
understand load balancers and firewalls and all the components
that software needs to live and run on.
Both are intrinsically important, and they are symbiotic. If business
is from Mars and IT is from Venus, then developers are from the
Milky Way, and operations are from another galaxy, far, far away.
Not only are the two teams far removed intellectually, they are
typically far removed geographically and organizationally as well.
Very often, they report into two completely different parts of the
business.
And so, the great war has raged, since the dawn of IT time.
Developers sniping at operations. Operations sniping back at
developers. These are two sets of individuals symbiotically linked,
yet in a constant state of tension and warfare. Frankly, it’s a miracle
that anything at all gets deployed into production today in many
organizations. When it does though, it typically takes weeks or
148

..MS2019..
 DevOps and application development

months of waiting, testing, deploying, fixing, re-deploying, more

waiting, more testing and so on and so on. Everybody is doing
their job the way they are supposed to be doing it. The one set of
people who suffer is the business.
business.

Figure 25 - The DevOps problem statement © Servian

Restructuring your team

Developers and operations need to work seamlessly together in this
new world. What does this mean to the way their teams are
structured and how they collaborate?
The answer is, a lot. As we stated above, DevOps is not a thing in
its own right. It is comprised of several initiatives and changes to
the way people work. Most importantly, it requires a change of
culture. There can be no more them and us. There can be no more
division and conflict.
Patrick Debois, the father of DevOps, coined the term at his first
conference. From then on organizations went about creating new
silos of teams that would do both the dev and the ops, or
somewhere in between and own the “pipeline” or delivery. This
was precisely what Patrick was looking to avoid.
At the heart of the DevOps machine and culture movement is the
quest to remove the number of handoffs between the teams and
the organizational structures involved in a task. That applies
whether it is a dev task or an operational one. Building a team that
149

..MS2019..
DevOps and application development

sat in between caused the problem to become worse. In more

modern DevOps practices, we’ve learnt that team responsibility
needs to remain with the team that is best qualified and skilled. So
how is the DevOps team established? It’s about lowering the walls
between the teams and providing a high level of communication.
This requires a common language. Teams still own their respective
areas of responsibility, but all teams have the ability to make change
in all other verticals.
The best practice example of this is that of the glorious “firewall
change”. The development team may well require a port changed.
But they may not understand the implications of just opening up
port 80 to the whole world from the database server. They ask the
operations team to do it, and there follows the customary fight.
The operational team would do better to allow the development
team to see the infrastructure as code that wraps the security port
protocols. That will allow the developer to “raise a pull request
with the change”.
This gives the operations team two things.
1. The opportunity to explain to the development team why
just opening the port may not be a good idea. They can
explain what else needs to be considered or changed.
2. The ability to protect against the developer just making the
change and causing issues.
Likewise, the operational team may want to look at the developer’s
code or at least the process flow. Then they can see why they’re
trying to make a call on port 80. They can talk to the developer and
suggest a different way to make the call. Either way, that
communication is the new “DevOps Team” at work. Roles and
responsibilities are respected, but hand-offs are reduced. If the port
change was fine to make, the operations team would just click go
and allow the roll out to occur – no questions asked.
There is no simpler practical step to achieve this than a realignment
of reporting structure. Today, operations typically report into some
kind of head of operations. Development reports into some kind
of head of development. Often these reporting lines don’t
converge until you get into the upper echelons of an organization
at some kind of directorial level. This doesn’t work. Within a
DevOps world, either developers need to report into operations, or
operations need to report into development. Or better still, do not
150

..MS2019..
 DevOps and application development

have operations and developers. Just have a software engineering

organization that consists of folks with a slant towards
development and folks with a slant towards deployment and
management. Some organizations completely drop the term
operations team and rename these people site reliability engineers.
This demonstrates that their role is now all about making code
work effectively and efficiently in production, and that they work
hand in hand with developers.
There is another substantial change to understand and act upon.
The roles of some of the people within each of these teams must
change. Developers need to start thinking a bit more like
operations folk. They should understand that they are not just
writing code that will end up on some server somewhere – a server
that they have no appreciation of or care for. They now know their
code will end up within some form of Cloud service, probably
chosen by either themselves or the architects within their team.
They need to make sure their code is designed in such a way that it
will seamlessly deploy into these services This requires that they
respect some form of service-oriented or microservice-based
architecture.
The bigger change is typically within the operations team. As we
will explore below, there are the cultural changes and adaption of
roles. But the biggest change is the tooling and automation that
DevOps necessitates. IT pros have been doing automation and
scripting for many years. In a DevOps world this takes on an ever-
greater importance. It becomes the only way that anything can ever
be done. No more manual steps. No more workarounds. No more
provisioning environments. Everything needs to be automated.
Everything becomes templated. This may prove a big challenge to
some team members who are used to doing things their way. They
are accustomed to drawing low level infrastructure design
documents. Some folks may not make it. Others will embrace the
new way of working and become excited about the power this level
of automation can bring.

151

..MS2019..
DevOps and application development

Automation and tooling

Good. You’ve fixed the culture. You’ve fixed the organizational
structure. Everyone is behind this new world and this new Digital
Operating Model. What’s next? Tooling.
Those of you running a mature software development function will
have some form of software development lifecycle management
(SDLC) tooling. You may use TFS. You may use Jira. You may use
several tools. They probably work quite well and help you manage
requirements capture and source control. Those functions remain,
and by and large will be unchanged in a DevOps landscape. What
you need to add is continuous integration and continuous
deployment (CI/CD).
There are, again, a vast variety of tools in this space It is possible
you may use some of them already. Examples are: Azure Pipelines,
Chef, Puppet, Ansible, Jenkins, TeamCity, and Octopus. There are
more tools than you can shake a stick at. Each has strengths, and
each has weaknesses. This book will not attempt to delve into the
details of how you set up an end-to-end CI/CD pipeline. We will
not discuss how you inject security and shift left thinking. Nor will
we go into which tool makes most sense based on what
technologies your developers use, and what technologies your
operations team feel most comfortable supporting. That would
require a book of its own. There are many books on the market
you can refer to. We will only introduce you to some of the
concepts. The important thing is for you to ask the smart questions
about how you can introduce DevOps into your organization.
Ultimately though, you need to do the research and make this
tooling the foundation of your Digital Operating Model. In a world
of digital disruption and digital innovation your application
development function becomes your powerhouse. Giving teams
access to the best tooling is the best way to help them be effective
and deliver the right outcomes for the business. Get it right and
you’ll truly be able to deliver the dream outcome: rapid, agile
delivery of software to your users and customers.
If you want to skip a step, however, and just choose something that
works quickly and simply out of the box, Azure DevOps is your
answer. Rather than cobbling together multiple different tools, pick
one integrated platform.
152

..MS2019..
 DevOps and application development

Immutability
From the Latin, “immutabilis”. Unchangeable.
In the introduction we called out a significant challenge. This is the
difficulty of supporting change within a traditional ITOM. We
made the point that change is difficult. Change is dangerous.
Change is something to be avoided. When we go deeper though,
we discover these statements hold true because of one thing. That
one thing is called mutability.
You may have heard about mutable and immutable – maybe in a
computer science class. It’s a programming thing, right? Correct.
But now it’s an infrastructure thing as well. As infrastructure
becomes programmable, so the concepts found in programming
become more applicable.
We’ve all experienced or heard stories about bugs and defects that
appear in production. These bugs weren’t noticed or couldn’t be
reproduced in UAT. The environments look exactly the same, yet
somehow there is some tiny difference that you just can’t quite
track down. There goes your weekend!
In today’s Mode 1 world, we have mutable instances of
infrastructure. They are ones we can mutate. Mutable is defined as:
‘that may be changed; subject to change’. The most basic example
of a mutation is patching. We don’t rebuild a server from scratch
each time a patch is released from an updated ISO which has the
patch baked in. We patch in place. But every time we install a patch
or deploy an updated version of an application that runs on that
server, we introduce tiny variants. Now we have tiny differences
between what may or may not have succeeded at an MSI installer
or DLL packager level. Two servers which started out identical are
no longer identical. If you update different configuration items
stored on these machines enough times, bad config or defects can
also creep in. Indeed, the vast majority of defects in production are
due to bad config across environments.
Why do we act in this way if it’s so prone to producing error? The
answer is because anything else is prohibitively expensive and
complex. Building a server, however automated it is, takes time. It
requires people. It needs coordination and approval. Applications
are typically deployed on to multiple servers, all of which need
provisioning. Applications won’t magically install themselves.
153

..MS2019..
DevOps and application development

Someone has to do it. When you reinstall something, it typically

breaks – at least 90 percent of the time.
Well, that used to be the case. But it’s not the case anymore. The
central concept behind Cloud computing is infrastructure and
application automation. This delivers us the ability to leverage
immutable environments. They are also known as ephemeral
environments. We can make use of Azure Resource Manager
(ARM) templates and technologies such as Desired State Config
(DSC) and Azure Deployment Manager. Additionally, we’ve seen
the vast adoption of Hashicorp Terraform and the HCL language
used to both drive and integrate with ARM templates to manage
resources in a more programmatic fashion. With their help an
automation specialist can describe exactly what an instance of an
application looks like. It doesn’t matter how many servers and/or
services it is comprised of. The automation specialist can describe
the components of the application and how each can scale up and
down independently of the other components. They can also
describe exactly how to deploy the application and the
configuration itself onto these underlying IaaS and PaaS services.
They no longer need to worry about patching (whether it's using
IaaS or PaaS). At a moment’s notice a completely new, ephemeral
copy of the environment and application can be recreated from
scratch. That’s game changing. It’s also organizationally changing.
We wrote earlier about the changes that need to be made within
your teams. Immutability is the central concept driving many of
these requirements for change. We no longer feed and water (that
is: administer) environments. We now create new ones through
scripting. The more automated we can make the deployment, the
better we can test it. At the same time, the likelihood lessens of
defects getting into production. That’s why these new folks are
typically called site reliability engineers. Their job is to stop defects
getting into production.
However, I can’t just take an application I have now and magically
make it support this new world. Infrastructure automation is very
similar to software automation and development. It takes time. It
requires new concepts to be introduced into existing architectures.
Concepts such as stateless operation and centralized configuration.
When we talk about applications which can exist in this new world,
we typically talk about Cloud-native applications. Almost
invariably, these applications are immutable. Proper DevOps also
154

..MS2019..
 DevOps and application development

requires immutable environments which can be automated and

deployed on demand.

Containers
To alleviate some of the challenges with IaaS-based deployments,
containers have emerged over recent years. They offer some
significant advantages.
A container is a bit like a VM. But instead of containing an entire
server operating system, it just contains the application files.
Multiple containers can then run inside a VM which itself includes
the underlying operating system. Scaling containers up and down is
quick and simple compared to scaling entire VMs up and down.
The industry standard container is Docker.
Containers by themselves are not especially helpful, though. In the
same way as you need something like Azure to automate and
marshal the creation and destruction of VMs, you need the same
for containers. You need high availability and fail-over between the
underlying VMs. You need orchestration.
There are several open source container orchestration solutions in
market. They include DCOS and Docker Swarm. The newest
entrant, and the one which has become the industry standard in the
last 12 months, is Kubernetes (K8s). K8s originated from Google
but is now widely adopted by the other two major Cloud vendors,
Microsoft and Amazon. Even VMWare, once in danger of letting
the Cloud world sail by, has gotten into its K8s groove. Google’s
multi-Cloud platform, Anthos, is also built on K8s. One of the
founding architects of K8s now works for Microsoft. Under his
guidance Microsoft has bet big on the Azure Kubernetes Service
(AKS). This fully managed platform service automates and
orchestrates all the work involved in deploying and managing
containers.
If you’re more focused on .NET, you can also make use of the
Azure Service Fabric. Service Fabric brings all the capabilities of
platforms such as Kubernetes, but also has significant additional
capabilities.

155

..MS2019..
DevOps and application development

Platform services
Whilst many believe containers are the only way to design and
build modern applications, this is not necessarily true. One of the
major advantages of working with a Cloud platform such as Azure
is the availability of platform services. These are typically referred
to as PaaS.
PaaS is a very different beast from infrastructure services or
containers. This is so even when you are leveraging immutability
and automation. At a high level, PaaS services are a set of
capabilities delivered by the underlying Cloud service which can be
consumed as services rather than delivered from servers – that is
servers you have to patch and maintain or containers you have to
deploy and orchestrate. Rather than building an SQL server cluster,
I can now consume SQL as a service. I simply pay for the storage
and performance requirements I want.
Naturally there are pros and cons to any deployment method. The
advantage of PaaS services is that users no longer need to maintain
underlying servers or container orchestration platforms. The
downside is users lose some of the control and customization
which is possible when you’re in control of the end-to-end stack.
Most platform services are themselves delivered under the covers
by containers. SQL Azure is built on Service Fabric. App Service
deploys IIS instances inside containers on top of VMs. The
platform service is, in effect, a container orchestration service for
specific containers – one that is designed and managed by the
Cloud vendor.
In a DevOps setting, PaaS services can have significant advantages.
Provisioning PaaS services is exponentially more straightforward
than automating the provisioning and configuration of VMs or
containers programmatically. Users can instantiate a complete
platform service with a handful of API calls. That is instead of a
complex set of deployment scripts. This makes creating immutable
environments much more straightforward. Users can create and
tear down specific PaaS instances more easily than a set of VMs or
containers which might support them.
Platform services can also be seamlessly scaled up and down based
on performance metrics. How many
transactions/messages/operations do I need to support per
156

..MS2019..
 DevOps and application development

second/minute/hour? The pricing model also exactly reflects these

performance-based characteristics. With a VM-centric view of the
world I might not know if it was more CPU, RAM or disk
performance which was required. You typically just scale all of
them up.
One relatively new entrant into the PaaS space is Functions-As-A-
Service (FaaS). FaaS is an additional abstraction on top of PaaS that
allows you to deploy and execute individual lines or modules of
code. The pricing model for FaaS is typically based on the number
of times code executes, rather than the size of the execution
environment (server or container). This model allows you to align
demand and cost even more closely. It also allows you even greater
ease of deployment. Within Azure, FaaS is delivered by Azure
Functions.
The downside of PaaS is that you do leave yourselves open to
some level of platform dependence. The dreaded lock-in. If you
target a specific platform service on a specific Cloud, it can be
more difficult to migrate to another Cloud vendor in the future.
Whilst lock-in is typically overstated by the multi-Cloud zealots, it
is nevertheless a real danger. Before you become too dependent on
a given service, understand how you might theoretically migrate
from one to another. For databases-as-a-service this might be as
simple as a backup and restore. For event-based services, this might
be a change of API call. If you want to use functions, choose a
language that works across the FaaS offering of multiple Clouds
(such as .NET Core). Don’t use lock-in as a reason not to use
platform services but be wise to what you’re getting into and how,
in theory, you might move if you needed to. In general, the ease of
consuming these services easily outweighs any remediation work
you might need to do in the theoretical scenario of having to move
in the future.
One of the safest places to explore and take advantage of native
FaaS services, like Azure Functions, is in the realm of operational
glue. That is, where services need to manage an event and trigger
another service or event, Azure functions play a perfect role. For
instance, suppose you wanted to instantiate a container that
processes the vectors in an image dropped into Azure blob storage.
You could produce an Azure Function that executes only when a
new file is realized in a particular folder. That would lead to a call
to the Azure Container Service to spin up a docker container and
157

..MS2019..
DevOps and application development

process the file for vectors. Finally, that information would be

passed back to a file in another folder. This can be quite powerful
when deploying a fully event-driven and automated micro and
Azure service architecture.
There are hundreds of platform services on Azure for data,
messaging, media, device management, identity, security, and much,
much more. It’s vital that you explore how PaaS services might
benefit your developers. Find out how PaaS can streamline your
application development workflow. Be bold. Mandate PaaS first.

Testing
Alongside immutability, a central tenet of DevOps is testing.
Immutability exists to aid, amongst other things, effective testing.
A key concept to understand is a failure model. There are two
kinds of failure model: a deterministic one and a non-deterministic
one. They are, in laymen-speak, a failure you can discover through
effective testing and one which you cannot. The more wildcards
you can remove from the equation (such as mutable environments
or physical hardware), the more unexpected failures you can
eliminate. We want to exist, as much as is possible, within a
deterministic realm. Then we can, at least in theory, remove all
defects through effective, complete and automated testing.
Doing effective, complete and automated testing is no mean feat.
We need to test a lot of stuff. We need to test code. We need to
test logic. We need to test UIs. We need to test integrations. The
primary reason for testing is to fix things. So, we need to engage
developers, at every stage. We need to shift left.
“The term “shift left” refers to a practice in software development in which
teams focus on quality, work on problem prevention instead of detection, and
begin testing earlier than ever before. The goal is to increase quality, shorten
long test cycles, and reduce the possibility of unpleasant surprises at the end
of the development cycle – or worse, in production. 28

In a shift-left world, we need to increase the lines of

communications across the teams. Particularly we need better
communication between the automation specialists in ops and the
developers. The shorter we can get the feedback, the more quickly
we can spot the defects and get them fixed.

28 https://devops.com/devops-shift-left-avoid-failure/
158

..MS2019..
 DevOps and application development

In the first way, there was one-way communication from devs to

ops. In the second way, there is basic return of information from
ops. Only in the third way can we truly shift-left. That is when
there is constant, automated and detailed communication between
development and operations.
This might consist of immediate and interactive telemetry that a
developer can access in development. It might be detailed stack
trace information from an environment closer to production. It
might be a Microsoft Teams notification that a build has failed. The
communication can come in many forms. The easier it is for a
developer to access this information, the quicker they can find the
issue, and the quicker they can fix it.
The testing we introduce, therefore, needs to be as automated as
possible. We need to make extensive use of unit tests which can be
executed and run at build time. We need to use ephemeral
environments, so that our end-to-end testing can run against exact
replicas of what our environment will look like in production.

Figure 26 - The first, second and third way of communications

We need to knit testing deep into our software development and

release cycle.
One of the reasons change has been dangerous in the past is
because it can introduce unexpected behavior. Our testing must be
as thorough and automated as it can be. Then we will only be
exposed, by and large, to deterministic failures. This allows change
159

..MS2019..
DevOps and application development

to be introduced at a much faster rate, without introducing

additional risk. This is the raison d’être of Mode 2, Cloud-native
application development and delivery. It is key to your digital
transformation. If you can be sure change isn’t going to break
something, you’re building on solid foundations.

Pipelines
Pulling all these concepts together is the pipeline. A pipeline is the
principal component of continuous integration and continuous
deployment (CI/CD). It is a workflow – a set of instructions for
the steps required to deploy an instance or a component of a given
solution.
The first step is to take the source code from the version control
repository. This code is then built according to the build scripts.

Figure 27 - A typical DevOps pipeline

Unit tests are then run within the compiled application and
services. Once the tests pass, an instance of the application and/or
services are deployed into an ephemeral environment. Further
automated integration and acceptance tests are run before the
environment is swapped/promoted into a production
environment. Simple, right?
Unfortunately, not quite so simple! The pipeline is the magic that
holds the whole DevOps world together. It’s the asset that reduces
hand-offs between teams. If your application is complex, your
pipeline might end up incredibly complex and elaborate. You will
typically have multiple pipelines for multiple applications or
multiple tiers within an application. Depending on the tooling you
use, your pipeline might need to speak to a plethora of separate
tools and processes. Anyone who has been involved in automation
knows that to automate a manual process is several orders of
160

..MS2019..
 DevOps and application development

magnitude more complex than you might imagine. Things always

find a way of going wrong.
But once you get your pipeline(s) working, it’s something to
celebrate. It’s the magic that allows you to increase your rate of
change exponentially. Do you need to change something? Just
modify the code, check it in, press a button. Ta-da! It’s in test.
Press another button, and, ta-da! It’s in production. When you can
do this, you embrace change rather than fear it. You might not hit
Amazon’s million releases a year. But you might manage a release
every week, or maybe even every day. That’s huge progress for the
vast majority of organizations.

Security
Security plays a very important part in this new DevOps world.
Historically, security probably got involved reviewing low level
designs. They may have asked for code drops to scan the source
code. They probably enforced security software / patching
schedules / penetration within production. It was a one-way
conversation. Security would say in effect: We’re going to tell you
what to do, and once you comply, we will allow things to happen.
In this new world, security can become embedded into the process
end-to-end. Policy is the key here. In the same way that we can
express and enforce policy within our infrastructure deployments,
we can now achieve the same within our application deployments.
Rather than stipulating requirements up-front, such as encryption
at rest, security folk can now access the same deployment scripts,
templates and policies as developers. Security folk can modify these
templates directly, enabling options and settings. When you
leverage native compliance toolsets such as Azure Policy, these
same controls can be enforced.
Security-related steps can also be inserted into the pipeline. The
security movement within the DevOps methodology and mind-set
has grown to such an extent that the term ‘DevSecOps’ has
emerged. Entire groups of individuals are emerging just to tackle
the automation that is security within the pipeline. Security used to
be something designed at the beginning and validated at the end
prior to the “go-live day”. but we’re now seeing organizations work
to trust the pipeline over the end environments.

161

..MS2019..
DevOps and application development

Automated code-scanning tools can check source code before it is

built to look for vulnerabilities or bad code. If you’re using
containers, these containers can be scanned for viruses at the point
they are created and sealed. You can then deploy as many instances
of each container as you like, safe in the knowledge that each is
virus free. You don’t need to check your entire environment any
more in production. You only need to check each of its building
blocks at build time.

Monitoring
In the last chapter we explored some of the logging and monitoring
tools available for your heritage estate. Depending on your
application architecture, these tools also have applicability in this
new world. We also have access to a range of other tools and
services which are designed for a world of custom software
engineering.
These tools capture additional metadata and telemetry which is not
typically available to more infrastructure-focused tooling. Within
Azure, Application Insights (part of Azure Monitor) provides rich
data about the functioning of applications at the code, module, or
container level. The following kinds of questions can be answered.
How long does this call to the database take? What is the spin-up
and spin down time of a container? What is the total number of
individual transactions which make up my higher-level
transactions? For instance, how many database calls are made per
order? Having deep visibility into the functioning of your
application is vital to ensuring its health. It also allows you to
optimize its cost footprint.

162

..MS2019..
 DevOps and application development

Figure 28 - Azure Application Insights

163

..MS2019..
DevOps and application development

164

..MS2019..
 The Technical Questions

Chapter

14
14The Technical Questions
Everyone hears only what he understands
Johann Wolfgang von Goethe (Writer, 1749-1832)

W E have outlined in this Part 2 of the book everything you

need to think about as you start to build out your Digital
Operating Model. As you have discovered, there is a lot
to think about. Now we ask a second set of questions so
that you can dig deeper. We want to position these challenges and
decisions by means of questions in tabular format.
These questions are presented under the same headings as the
preceding chapters (8–13) of Part 2 of this book. We pose 20
questions under each heading. If you need some context around
the questions, turn back to the relevant chapter. This will remind
you about the rationale of the question and will guide you towards
the best answer.
You may not find the answer to every question. You might need to
consult internally. You may need help from your vendors or from a
partner. Don’t panic. This is a complex subject and answering all of
the questions will take time.
 Chapter 8 – Strategy and Service Providers
 Chapter 9 – Procurement and Financial Governance
 Chapter 10 – Service Management
 Chapter 11 – Access Control, Security and Policy
 Chapter 12 – Monitoring, Management and Automation
 Chapter 13 – DevOps and Application Development

165

..MS2019..
The Technical Questions > Strategy and Service Providers

14.1 Strategy and Service Providers

Getting your Cloud strategy right from the outset is the most
important factor in determining the success or failure of your new
Digital Operating Model.
In chapter 8 our headings are:
 Setting off in the right direction
 Communities of practice
 Multi-Cloud
 Hybrid Cloud
 Your estate
 Bimodal IT
 Service provider relationship
 Service level agreements and
 Support
Some questions to ask about these topics follow in the table below.

166

..MS2019..
 The Technical Questions > Strategy and Service Providers

 Question Why this matters

⎕ 14.1.1 How We discussed in great detail in Part 1, the
much does the business part of this book, how IT can be
business think IT used as a strategic differentiator. It’s vital
can be used as a that the business accepts this as true. It’s
key differentiator vital for them to accept the part you need
in the market? to play in transforming the business.
⎕ 14.1.2 Has the This is a very important starting point. If
business got the business is struggling to articulate what
clear and it needs to do to be successful, it is very
documented difficult for IT to draw alongside and
goals in the form deliver meaningful change.
of a business
strategy with a
1–, 3–, and 5–
year view?
⎕ 14.1.3 What is Part 1 examines this topic in more detail.
the perception of The key here is that if IT is relegated to the
the IT function role of keeping the lights on, it will take
within the some effort to win hearts and minds. It will
business? be hard for IT's role to be elevated to that
of trusted advisor to the business.
⎕ 14.1.4 Has the This is where real traction starts to take
business place. The objective here is to map
undertaken an business objectives to IT objectives. The
exercise to align purpose is to ensure that the IT function is
IT strategy with delivering the capability to the business in
the business the way and at the pace they need for its
strategy? success.

167

..MS2019..
The Technical Questions > Strategy and Service Providers

 Question Why this matters

⎕ 14.1.5 Can the If the business has a strategy and the IT
business group has a strategy and they are aligned,
articulate what you are in a good place. You will be able to
IT delivery and deliver on the vision. However, the
transformation business stakeholders must understand this
programs are plan and buy in to its importance. It is
aligned to which good to have the plans; it is better if
business goal? everyone keeps them front of mind and is
Does it know engaged in actively using and adjusting
how the IT them over time.
capability will
contribute to the
stated goal?
⎕ 14.1.6 If the If none of this exists in your organization,
answers to it is important to get commitment to
questions 16.1.4 undertake some of this planning work.
and .5 are no, This is something you can do internally.
would the Alternatively, there are external
business be consultancies who can help with this
prepared to important piece of work.
undertake a body
of work to
understand and
document these
objectives?
⎕ 14.1.7 Do Unless and until IT leaders accept their
senior IT importance to driving home the businesses
stakeholders in strategic vision, they will remain relegated
the business see to a supporting role. If you don’t have
their role as a key people in your IT team who accept this
contributor to responsibility, get some new people.
the overall
business success?

168

..MS2019..
 The Technical Questions > Strategy and Service Providers

 Question Why this matters

⎕ 14.1.8 Is the Traditional outsource contracts seek to
business locked avoid change and maintain the status quo.
into a multi-year They are incompatible with a DOM. If you
outsource? If so, are currently outsourced, find out what you
what options are can do outside of this relationship. Start to
there to deliver gear the business up for the next renewal.
services At that time, you can insource again or
alongside this? engage a modern Cloud solutions provider.
⎕ 14.1.9 Does One of the major benefits of moving to a
your Cloud model is visibility into costs. You
organization use can cross-charge other departments for
charge back for their IT. If these departments have not had
IT services or do to pay for their IT before, they may not see
IT hold all the IT this as a welcome step!
budgets?
⎕ 14.1.10 Do you Multi-Cloud sounds great. Why wouldn’t
have a vision for you want it? In reality, it’s orders of
a multi-Cloud magnitude more complex than you might
strategy? realize. Also, it doesn’t necessarily deliver
huge benefits. It’s a decision not to be
taken lightly.
⎕ 14.1.11 What is If you do want to go multi-Cloud, what is
driving your the driving factor behind this decision? Is it
decision to go cost optimization? Is it avoiding lock-in? Is
multi-Cloud? it to leverage capabilities unique to a
specific Cloud platform? Sometimes there
are valid reasons. If you really want to go
multi-Cloud, make sure you can give clear
explanations for wanting to do so. You’ll
need them later when you have to justify
the huge additional complexity.
⎕ 14.1.12 If you do There are strategies that allow you to hedge
want to go multi- against the additional risk and complexity
Cloud, can you of multi-Cloud. Choosing a development
hedge the platform consistent across all container
complexity risk? platforms can help. Kubernetes is a good
example. Also consider the concept of
general-purpose vs specialized Clouds.
169

..MS2019..
The Technical Questions > Strategy and Service Providers

 Question Why this matters

⎕ 14.1.13 If you do Monitoring and management are as
want to go multi- important as deployment in a multi-Cloud
Cloud can you world. Choose a solution such as Azure
leverage Log Analytics which works across different
common Clouds.
tooling?
⎕ 14.1.14 Do you There are several reasons why your
have legacy applications might not move to the public
applications Cloud. Your apps might run on legacy
which might not hardware such as mainframes or specialist
move to the machines. You might not be able to license
public Cloud? your applications within a virtualized
world. You may have a very low latency
requirement between different tiers of an
application. In these situations, you need to
consider a hybrid Cloud strategy.
⎕ 14.1.15 Have you There are reasons why organizations might
just made have done this. Your hardware might have
significant been so old and so unreliable that you had
investments in no choice. You might have been led to
replacement believe that a private Cloud was the same
hardware? as a public Cloud. You might not have
discovered Cloud early enough. In these
circumstances, there can be challenges to
adopting public Cloud. But in a hybrid
world you might still be able to take
advantages of some public Cloud
capabilities.

170

..MS2019..
 The Technical Questions > Strategy and Service Providers

 Question Why this matters

⎕ 14.1.16 Are your Many organizations, especially public
requirements for sector ones, have rules about what data can
data storage be stored where. Official and official
incompatible sensitive data can be stored in the public
with public Cloud, but secret data cannot. In this
Cloud security? circumstance, hybrid can deliver some
major advantages. You can construct a
DOM which straddles public and on-
premises Cloud. At the same time,
dedicated hardware in a traditional
datacenter will look after high-security
requirements. Appliances such as Azure
Stack can help here.
⎕ 14.1.17 Do you Hybrid can play an important role here.
store large Frequently accessed, hot data can remain
quantities of on premises. Historical, infrequently
historical data? accessed data can be moved to the public
Cloud.
⎕ 14.1.18 Do you Hybrid can also play an important role
need a disaster here. Your primary environment will
recovery remain on-premises, but your fail-over,
solution? disaster recovery environment moves to
the public Cloud.
⎕ 14.1.19 Do you VMWare was the gold standard in days
run VMWare on gone by. But it has failed to keep up with
premises today? the industry’s move to public Cloud. Its
aborted attempt to get into this world with
vCloud Air was mothballed. If you run
VMWare on premises today, you will
struggle to realize a true Cloud-native
model. VMWare on Azure has just been
released which will at least offer some
hybrid capabilities.

171

..MS2019..
The Technical Questions > Strategy and Service Providers

 Question Why this matters

⎕ 14.1.20 Are you a Work through your network connectivity
global plans before you start deploying workloads.
organization with You may end up with multiple hybrid data
both centralized centers in region with global network
and regional connectivity. Microsoft has ExpressRoute
deployments of premium which will allow you to use their
workloads? global network instead of your traditional,
expensive WAN provider.

172

..MS2019..
 The Technical Questions > Procurement and Financial Governance

14.2 Procurement and Financial

Governance
This section corresponds to chapter 9 of the book. Once you’ve
decided upon your Cloud strategy, you need to build processes
around procuring it and governing it.
The way you procure Cloud services will be very different from the
way you procured on-premises software. The organizations you
procure it from will likely be different. The questions you ask of
providers will almost certainly change.
The way you allocate and manage the cost of your new Cloud
environment will change fundamentally. Charges will change from
fixed up-front costs to variable ongoing costs. You could spend
more than your thought. But you have the opportunity, with work
and attention, to spend less than you thought. These challenges and
opportunities need to be understood and managed.

173

..MS2019..
The Technical Questions > Procurement and Financial Governance

 Question Why this matters

⎕ 14.2.1 How You may already have a sophisticated
does your mechanism for understanding IT cost
business account today, based on cost per user numbers and
for IT spend at cost per line of business application.
the moment? Alternatively, you may have a few big lines
in your P&L and a lack of structure behind
this.
⎕ 14.2.2 Can you If you can’t do this today, start to think
state your IT about how you might represent this in your
costs per user or procurement and IT accounting costs.
per application? These will start to be the pillars you build
on moving forwards.
⎕ 14.2.3 Do you Whilst the quarter end and year end pattern
normally wait will never completely disappear from
until year end to vendors, with a consumption-based model
bag the best deal and an inability for the vendor to book
from vendors? revenue in line with orders, steep discounts
to plug a gap in their numbers don’t
typically exist within a cloud world.
⎕ 14.2.4 Where Historically, most IT license spend was
do you buy your with IT resellers or License Service
IT licenses and Providers (LSPs). In this new world,
services from licenses can now be procured from a
today? number of different people and styles of
organizations. Cloud managed services
businesses (MSPs) can bundles
professional services and cloud
consumption into one bill.

174

..MS2019..
 The Technical Questions > Procurement and Financial Governance

 Question Why this matters

⎕ 14.2.5 Do you Whilst there are still some discounts
expect large associated with large commitments,
discounts from typically for SaaS solutions, with IaaS and
vendors? PaaS platforms, the margins are
substantially lower and there is less
opportunity for the vendor to offer these
discounts. Instead, often vendors will
leverage investment funds to pay for
workloads to be moved to the cloud to
drive increased usage and consumption.
⎕ 14.2.6 Have you Your procurement pipeline is an important
chosen asset to allow you to work with multiple
Microsoft as one cloud providers, you will typically have a
of your strategic different process to engage with your
cloud partners? strategic partners. If you’ve already chosen
Microsoft, don’t let setting up a
procurement pipeline for other vendors
stop you getting to work with your
strategic partners.
⎕ 14.2.7 Do you Many traditional RFP sets of questions
have a set of don’t fit in this new world. You’ll need to
procurement review them and work towards a new set
questions that that are fit for purpose. We have included a
are fit for a cloud sample of the kinds of questions you might
world? ask below.
⎕ 14.2.8 Where is Many SaaS vendors use a cloud
your SaaS deployment model themselves. If they use
application one of the major cloud providers, ideally
delivered from? one you use yourselves, you can forgo
many of the detailed hosting questions you
may have asked in the past as you will
already know the answers to them.

175

..MS2019..
The Technical Questions > Procurement and Financial Governance

 Question Why this matters

⎕ 14.2.9 Do you There is no right or wrong answer to this,
use an IaaS or but typically an IaaS deployment model
PaaS deployment may suggest a legacy application which has
model? been “cloud washed”. Vendors using PaaS
will have needed to have done extensive re-
architecting which demonstrates their real
commitment to cloud.
⎕ 14.2.10 What is Firstly, they should immediately understand
your shared this question. If they don’t panic. Cloud
responsibility vendors support SaaS vendors to some
model with your extent, depending on whether it’s IaaS or
cloud provider? PaaS. SaaS vendors should know intimately
what they are responsible for and what the
cloud vendor is responsible for and should
have a documented RACI matrix for this.
⎕ 14.2.11 What is A cloud deployment model doesn’t
your high magically give SaaS providers HA/DR out
availability/ of the box. These concepts are as
disaster recovery important in a cloud deployment model as
(HA/DR) plan? an on premises one. Make sure your SaaS
vendor has thought through all these
challenges and has a good answer to them.
⎕ 14.2.12 What is When you move to a SaaS-based
your SLA? application you are going to be wholly
reliant on the SLA of your SaaS provider.
How does their SLA intersect with that of
their hosting provider? Have they carefully
thought through composite SLAs based on
the range of cloud services their application
hosting leverages?

176

..MS2019..
 The Technical Questions > Procurement and Financial Governance

 Question Why this matters

⎕ 14.2.13 What Whilst less important due to the nature of
technology stack SaaS delivery it’s always interesting to know
do you use? what technologies the SaaS provider uses.
What is your Do they use a traditional relational
architecture? database or a schema-less NOSQL
database? Do they use SOAP or REST
service endpoints? All of these will give
you clues about how serious and
technically sophisticated the vendor is.
⎕ 14.2.14 Is your Again, there is no right or wrong answer to
solution single this. Vendors might choose single-tenancy
tenanted or for compliance or performance reasons,
multi-tenanted? alternatively they might choose it because
they don’t want to invest the time to re-
architect for the cloud. Typically, a multi-
tenanted design is better as it reduces the
vendor’s cost, which they can pass on to
you, and it demonstrates it is actually SaaS
software, not just on premises software
“delivered through the cloud”.
⎕ 14.2.15 What is True SaaS vendors will operate on an “ever
your update green” basis and make frequent changes to
frequency? their code base. Those that only deliver
updates every few months probably
haven’t fully embraced cloud / DevOps
within their own organization.
⎕ 14.2.16 Does all Traditionally, when there were large
IT spend need to CAPEX spends behind infrastructure
go through projects, this approach made sense. With
procurement? smaller procurements from cloud vendors,
this model can become prohibitively time
consuming and expensive.

177

..MS2019..
The Technical Questions > Procurement and Financial Governance

 Question Why this matters

⎕ 14.2.17 How are With cloud deployment models,
IT budgets configuration changes or adding and
expressed and removing users can have substantial and
spent today? immediate effect on your usage charges.
Budgets need to become much more
flexible and capable of being “spent”
without as much control and governance as
you may have today. Techniques such as
quotas rather than budgets can make more
sense.
⎕ 14.2.18 Do you Much of the governance you will introduce
have the ability to your cloud environment will be the
to apply budgets enforcing of cost allocation. In order to
to departments enforce this allocation, you need to have
or cost centers? the appropriate level of granularity within
your existing cost code structure.
⎕ 14.2.19 Do you By using cloud, you have fine grained
have approach control of who, at what level, can spend
levels of what. This is something you may consider
delegation within within your budgeting and financial
your budget and governance model. For example, a person
cost code in role x can increase spend by up to £x /
structure? hour, but someone lower down in the
organization in position y can only increase
costs by £y / hour. This allows a shift from
an absolute to a relative definition of a
budget.

178

..MS2019..
 The Technical Questions > Procurement and Financial Governance

 Question Why this matters

⎕ 14.2.20 Who is No matter how hard you try and what
now in charge of controls you put in place, you will almost
monitoring certainly end up spending more than you
spend and should. There will be things left on or
introducing licenses connected to people that don’t
optimization? need them. There will be servers deployed
that are too big or not being used. There
may be defects which drive increased
consumption. As you move to the cloud,
place someone in charge of cost
optimization. Someone whose entire job
revolves around asking all the annoying
questions about why this is needed and
why the specification of a given server is
too high. Give them tools to help them do
this programmatically and automatically.

179

..MS2019..
The Technical Questions > Service Management

14.3 Service Management

This section corresponds to chapter 10 of the book. Service
management is at the core of your IT operating model. It governs
just about everything related to how you consume and manage
your IT services.
You almost certainly have a slick process today. But your service
management realm will likely need extensive attention and change
as you move into a Digital Operating Model.
ITIL has served us all well over the years, but the prescriptive
processes in its current version may not be your bedrock into the
future.

180

..MS2019..
 The Technical Questions > Service Management

 Question Why this matters

⎕ 14.3.1 Do you ITIL offers a lot of value to organizations
use ITIL today? today and has seen a substantial update
with v4. But in previous incarnations, it can
introduce additional complexity and
processes which can severely reduce the
effectiveness of your DOM rollout.
⎕ 14.3.2 Do you Service Integration and Management
use SIAM today? (SIAM) was a great concept in its time. It
allowed large, monolithic and complex IT
operating environments to be broken up
into bite-sized chunks and distributed to
different service organizations. The idea
behind it is still valid, its current
implementation is not hugely appropriate
for a next generation DOM.
⎕ 14.3.3 Are you You may well have contracted with a SIAM
locked into a provider, or have SIAM constructs
SIAM model? embedded within your existing operating
model. If you do, it doesn’t mean you can’t
introduce a DOM. But if you do, you will
almost certainly have to free it from the
bounds and controls that your other towers
need to adhere to.
⎕ 14.3.4 Do you For organizations that have been
want to bring completely outsourced, bringing additional
additional partners in can be challenging. Consider
partners into a how you will manage these additional
multi-source partners alongside your incumbent. You
model? may need to introduce a role or team
responsible for vendor management and
integration.

181

..MS2019..
The Technical Questions > Service Management

 Question Why this matters

⎕ 14.3.5 How will Whenever you have multiple partners or
you manage vendors there are extra complexities
integration around how these partners and vendors
between these will work together. Ensure you spend the
partners? time to map out the interdependencies and
the processes you will use to hand off
between these people. This must be done
in a controlled way.
⎕ 14.3.6 Are you Portfolio is the more modern approach to
able to break IT service delivery. Rather than individual
your IT service service lines, think about how these can be
provision into a grouped-up into cross-tower functional
portfolio capabilities which can be presented back
approach? into the business.
⎕ 14.3.7 Do you You do? Great. You typically won’t need to
have a mature make major changes to this. Incidents will
incident still occur in the Cloud. You’ll need to
management manage them. Some will be minor; some
process? will be major. Understand how you will
interface and get updates from the new
players in your process. But for the most
part we’re talking evolution not revolution.
⎕ 14.3.8 Can you With a move to the Cloud there are
automate any of growing opportunities to automate things.
your incident Incident management is a candidate ripe
management for automation. Can you automatically
processes? undertake remediation work to fix a known
problem? Can you apply AI and machine
learning to automatically make suggestions
to support operators on things that they
can try?

182

..MS2019..
 The Technical Questions > Service Management

 Question Why this matters

⎕ 14.3.9 Can you With increasing automation and
apply a new standardization, there could be less need to
mindset in which actually fix things. In the same way you just
you destroy and throw away a broken TV and buy a new
recreate rather one, so will you be able to with aspects of
than fix? your IT environment in the future. If a
laptop is misbehaving, factory reset it and
automatically rebuild it rather than trying to
fix the problem. The same can be true for
servers and application environments.
⎕ 14.3.10 How Change management will be the process
mature is your most profoundly impacted by a move to
change the Cloud. Change management is in the
management front line of the battle between agility and
process? control. It is purposely designed to reduce
risk which in turn means resisting change.
Map out your change process and
understand how you need to go about
changing it (no pun intended!).
⎕ 14.3.11 Do you CABs generally serve two purposes. Firstly,
require all to inform everyone about what is
changes to go to happening and secondly to allow people to
a change speak up if they identify risk as part of a
advisory board change. The first of these can now be
(CAB) achieved by other means, such as release
notes and automated communication. The
second should become less of an issue as
change becomes the norm rather than the
exception. Review what changes really need
to go to a CAB and which can bypass it
moving forwards. This assumes that
appropriate controls, policies and guard
rails are in place.
⎕ 14.3.12 How will The evergreen nature of Cloud services
you deal with means new features will appear
authorizing new continuously. You need to understand how
capabilities on you will deal with them and how you will
the platform? release new capabilities to users.
183

..MS2019..
The Technical Questions > Service Management

 Question Why this matters

⎕ 14.3.13 Do you Successful organizations select a cross
have early section of their user population to act as a
adopters / test bed for new capabilities. Do this
champions before the capabilities are rolled out to the
internally? wider user group. In this model you can
undertake proofs of concept and detailed
testing.
⎕ 14.3.14 Do you As you move further up the Cloud IaaS >
have tiered PaaS > SaaS stack, there will be a
support teams diminishing requirement for in-house
across each of second- and third-line support teams. The
your service Cloud vendor will provide these. Think
lines? about how you can redeploy these
capabilities internally for instance into
automation teams.
⎕ 14.3.15 How will You may need support from your Cloud
you interface vendors, either in connection with service
with your Cloud or configuration issues. You must
vendor’s support understand who will deliver this support
organization? and how it will be delivered. Will you
interface directly with the vendor? Or will
you go via a reseller, or your Cloud
managed services provider? Understand
the roles and responsibilities in the chain
and what you will need to do in the future.
⎕ 14.3.16 Do you You will need to consider which
have a Microsoft participants hold which support contracts
Premier Support for your DOM. If you are working with a
Agreement? Cloud managed service provider who has a
Microsoft Premier contract and can make it
available to you, do you need your own one
moving forwards?

184

..MS2019..
 The Technical Questions > Service Management

 Question Why this matters

⎕ 14.3.17 How will As you move into a world of automation
you support and DevOps your support organization
internally will change profoundly. Issues impacting
developed service will move from being
applications? predominantly infrastructure or
configuration issues to being
predominantly code and automation issues.
Think about how your platform services
and development teams will play a part in
the support organization. ‘It worked in test’
will not cut it moving forwards.
Developers and automation specialists
would be wise to get ready for that call at
3am.
⎕ 14.3.18 Do you You may already have applications and
have shadow IT services which sit outside your service
in place today management realm. You need to consider
outside your how to bring them into this new service
control? management structure.
⎕ 14.3.19 Do you SLAs change radically as you move into
have clear and this Cloud world. You need to understand
measurable how you can meet your existing SLAs. You
business and must also manage the SLAs that come with
technology SLAs the matrix of new suppliers and vendors
in place? that form part of yourDOM. You may
consider introducing new measure and
KPIs such as Net Promoter Score (NPS) to
get a better understanding of the success
and operation of your new DOM?
⎕ 14.3.20 Are you User Level Agreements (ULAs) and Value
familiar with Level Agreements (VLAs) are the direction
ULAs and VLAs of travel for the interface between business
and IT. You need to understand them as
concepts and build a path to deliver against
these new metrics.

185

..MS2019..
The Technical Questions > Access Control, Security and Policy

14.4 Access Control, Security and Policy

This section corresponds to chapter 11 of the book. Getting access
to and securing your new Cloud environment is of paramount
importance. Security used to be the big block to moving to the
Cloud. That situation has now been reversed. Cloud vendors lead
with the security of their offerings. Many organizations still do not
appreciate that fact.
Using a secure Cloud platform is only half the solution. There’s no
point having a state-of-the-art security on your house or car if you
leave them unlocked! Security in the Cloud is a shared model. The
vendor is responsible for a great deal. But you’ll also be responsible
for a fair bit.
This challenge must be viewed through multiple lenses. Identity
and access control are the things to think about. So too is physical
and network security. Finally, you’ll need to think about how you
provide these capabilities to users and make sure your provisioning
and JML processes are secure.

186

..MS2019..
 The Technical Questions > Access Control, Security and Policy

 Question Why this matters

⎕ 14.4.1 Do you If you do, get them involved right now. Put
have a Chief the book down. Send them an e-mail. Tell
Information them you want to meet as soon as possible
Security Officer? to get them on board with what you are
thinking. If you don’t, you’ll regret it later.
They can be your greatest ally or fiercest
opponent. Find out which as soon as you
can.
⎕ 14.4.2 What is CISOs come from different backgrounds.
your CISO’s Among them you will encounter
approach to risk? fundamentally different approaches to risk.
Some will be open-minded. Others will be
set in their ways. If they love to look at low
level infrastructure designs, you’re going to
need to help them see beyond these. The
quicker you can identify how supportive
they will be with your Cloud journey, the
sooner you can find out how much time
you need to set aside to get them on board.
Without them, you’re dead in the water.
⎕ 14.4.3 Does If you do, this may be a good thing, or it
your may be a bad thing. If the security
organization principles and rules run to hundreds of
have a pages, you’re in for a lot of work. But at
documented set least you have a strong foundation to build
of security on. Get hold of the document and read it.
principles and Get a sense of how specific it is. In other
rules? words, determine to what extent it is
related to traditional methods and models.
It may just describe high level principles
and rules. If it simply describes best
practices, you’re good to go as you can
easily replicate these within a Cloud
context. If it mandates things like VPNs
and physical access tokens, you’re going to
need to get it changed, fast.

187

..MS2019..
The Technical Questions > Access Control, Security and Policy

 Question Why this matters

⎕ 14.4.4 Do you Your Cloud identity model, typically, is
use Active built directly from your on-premises Active
Directory on Directory environment.
premises today?
⎕ 14.4.5 What Moving to a Cloud identity model means
state is your your on-premises environment must be in
Active Directory good shape. There are things that you
in today? might have gotten away with on premises.
They include duplicate namespaces and
accounts with rogue characters. This kind
of thing won’t cut the mustard in the
Cloud. You may have acquired other
organizations and bolted their ADs onto
yours. Before you do anything, get your
existing environment checked over. Get a
report on how it looks and what you need
to do.
⎕ 14.4.6 Do you Physical access tokens were great. They
use physical served a purpose in their time. With the
access tokens Cloud, there are newer and better
today? technologies to use, such as Azure MFA.
These technologies are much more cost
effective. They allow additional forms of
authentication, such as phone calls, text
messages and mobile apps. They can even
stretch to the physical proximity of the
phone to the laptop or to a geo location.
⎕ 14.4.7 Do you You may leverage Active Directory
have an on- Federation Services (ADFS) to provide
premises single single sign-on to users. If you do, you will
sign-on solution probably have several connections you set
in place? up manually, exchanging certificates and
keys. It probably works OK most of the
time. You can probably keep this in place
for some things. But in time you’ll
probably want to migrate it to something a
bit more up-to-date such as Azure Active
Directory.

188

..MS2019..
 The Technical Questions > Access Control, Security and Policy

 Question Why this matters

⎕ 14.4.8 Do you There are Cloud-based single sign-on
have a Cloud- solutions in market. Okta is one of the best
based single sign- known. These solutions do a great job and
on solution? if you use one you will already be
experiencing many of the benefits of
moving to a new Cloud IdP. But if you do
have these solutions in place, find out
about the capabilities available natively on
Azure, such as Azure Active Directory.
You may not need these third-party
solutions moving forwards.
⎕ 14.4.9 How are JML in a Cloud world is very different
licenses impacted from how it is in the on-premises world.
for JML in a Providing access and revoking access to
Cloud world? Cloud-based services starts and stops cost
being accrued in near real time. It also
requires licenses to be available in near real
time. If you're on an EA, ensure you have
sufficient headroom to support bringing
users online quickly. If you are on CSP you
must fully understand the mechanism to
get additional licenses.
⎕ 14.4.10 How Single sign-on is great. But it does mean
does security users have access to a great deal of data
change for JML and different services wherever they are.
in a Cloud You need to ensure that your JML process
world? is quick enough. You must be able to
switch off access to these third-party
systems quickly as employees change roles
or leave the organization.
⎕ 14.4.11 What You probably use one of the big brand
security solution security solutions on premises today, such
do you use on as Check Point, F5 or Barracuda.
premises? Understand what the vendor’s approach is
to public Cloud. Do they support it? What
is their licensing and deployment model?

189

..MS2019..
The Technical Questions > Access Control, Security and Policy

 Question Why this matters

⎕ 14.4.12 Do you The answer to this is almost certainly yes.
require dedicated So, you need to understand what options
connectivity to are available from your telecoms provider.
your on-premises Do they support ExpressRoute? Which
environment? flavor? Who supports what?
⎕ 14.4.13 Do you The immediate answer to this is often yes.
want to route all And “yes” is sometimes OK. But
your Cloud understand the implications of this from a
traffic, including cost and performance perspective. Does
internet everything have to transit via this route, or
breakout, on can you separate out some traffic, such as
premises? Microsoft Updates?
⎕ 14.4.14 Do you Azure and ExpressRoute have very
need to inspect sophisticated controls for data routing, and
and control the access control. But they have limited
ingress and capabilities for traffic inspection. You may
egress of traffic need to consider additional security
to the Cloud? products and services to inspect this traffic.
⎕ 14.4.15 What are There are many approaches today to
your policies segregate traffic within internal networks,
today on internal including VLANs. Understand how these
network security? map to the Cloud. Moving to the Cloud
can also be a good excuse to ramp up your
internal network security controls as they
manifest within the Cloud. It’s much
simpler now to enforce stricter access
control rules between network segments or
require additional controls such as IPSEC.
⎕ 14.4.16 Do you If you don’t use these on premises today,
use host-based explore whether you can introduce them as
firewalls and part of your move to the Cloud. Best
other controls? practice now is to have as much security
within your networks as is enforced
between and into your networks.

190

..MS2019..
 The Technical Questions > Access Control, Security and Policy

 Question Why this matters

⎕ 14.4.17 Do you Carefully consider how a traditional service
employ a service catalogue fits into your newDOM. A
catalogue today? service catalogue can be useful for some
aspects of provisioning. In other ways it
can be constricting.
⎕ 14.4.18 Can you A Cloud environment can be a more useful
embrace the construct to dovetail into a service
notion of Cloud catalogue. It allows users to serve
environments? themselves through your own service
portal. But they can still use the native
capabilities, such as the Azure Portal and
APIs, to provision individual capabilities
and services.
⎕ 14.4.19 What There is a lot of information you could
information collect from users as part of a provisioning
would you need process. Try to balance the need for
to collect from comprehensive reporting with making the
your users at the process easy and streamlined.
point of
provisioning a
Cloud
environment?
⎕ 14.4.20 Do you Azure Policy is specific to Azure. It is also
have a policed by Azure. Some organizations may
requirement to want to enforce policies across multiple
span policy Clouds. Or they might want to provide a
across multiple separate, independent view of their
Clouds or compliance. Third party tools can provide
provide additional capabilities in this space.
additional Examples include: Palo Alto Evident,
reporting on NeuVector and DivvyCloud.
compliance?

191

..MS2019..
The Technical Questions > Monitoring, Management and Automation

14.5 Monitoring, Management and

Automation
Monitoring, management and automation are at the heart of your
Digital Operating Model. In chapter 12 we focused on managing
your estate, as it looks today, in the Cloud. Azure provides
capabilities to assist with automating your heritage workloads as
you go through a process to make them more Cloud-native.
We described this model as a Gartner Mode 1.5. That is
somewhere between a legacy, on- premises environment (Mode 1)
and a brand new, Cloud-native Mode 2.
Moving to a more modern approach to monitoring and managing
heritage workloads can have a big impact on your ITOM. It will
almost certainly require migrating to Cloud-native tooling. That can
be a major undertaking.
With these questions we’ll hopefully help you to understand this
first stage of your transition to becoming Cloud-native.

192

..MS2019..
 The Technical Questions > Monitoring, Management and Automation

 Question Why this matters

⎕ 14.5.1 What There will probably be only a handful of
tool do you use people expert on your monitoring and
for monitoring management operating model. It’s
and managing probably something that’s been there
your forever. It just sits in the background doing
environment its thing. Find out who’s your expert and
today? do some research.
⎕ 14.5.2 Do you You may have tried migrating tools in the
have one, or a past. You may have a mix of different
number of tools, doing different things, supplied by
different tools? different vendors. Your tool for
monitoring patches may be different from
your tool for monitoring desktops.
Understand the entire landscape of your
toolchain. Ideally draw it up into a diagram,
if you don’t have one already.
⎕ 14.5.3 Do you If you do, great. Your job will be easier. If
use System you don’t it’s not a massive problem. You
Center? will probably need to migrate to new
tooling anyway. You may just have a bit
more work to do.
⎕ 14.5.4 Is your Many of the tools in market today
current (including System Center) were built in the
monitoring days of on premises. The clue lies in the
system designed name of some of them – names such as
for an on- LanGuard. These tools naturally continue
premises world? to evolve and support more Cloud
workloads. But it’s often very difficult to
use these tools to support your new Cloud-
native monitoring and your management
operating model.
⎕ 14.5.5 What In the same way that monitoring, and
tool do you use management tools were designed for the
for backup on-premises world, so too were backup
today? solutions. You need to consider whether
your solution is still valid in a Cloud world.
Could you leverage Cloud-native tooling?
You may be able to utilize a hybrid, where
your existing solution hands off long-term
storage to the Cloud.
193

..MS2019..
The Technical Questions > Monitoring, Management and Automation

 Question Why this matters

⎕ 14.5.6 How do In a server world, the process is largely
you do server similar. But now you can make greater use
patching today? of automation to power the end-to-end
patch cycle across complex architectures.
Azure can also automatically patch IaaS
machines if you enable this
⎕ 14.5.7 How do In the device world, with modern
you do device management, clients now generally take
patching today? their updates directly from Microsoft,
governed by Intune. Most users have these
updates forced on them as part of the
evergreen Windows as a Service.
⎕ 14.5.8 Are your With modern device management, devices
devices primarily become Cloud-enrolled rather than on-
on site or premises. This can have advantages, such
remote? as better support for remote working. But
it can have disadvantages, such as the
additional complexity of managing a large
number of office-based devices. Carefully
plan your hybrid device management
solution.
⎕ 14.5.9 How In this new world of modern device
complex is your management, Intune and MDM replace
AD GPO Active Directory and GPOs. Depending
structure? on the complexity of the policies you
enforce today, there could be substantial
work involved in migrating policies to
MDM.
⎕ 14.5.10 Which Many of the advanced, modern
edition of management capabilities are only available
Windows 10 do in Enterprise. You may need to reconsider
you use today? your licensing strategy for devices.
Enterprise is now available under the CSP
licensing program (it used to be just EA).
⎕ 14.5.11 How do With modern management, gold images are
you use gold typically no longer required. Devices
images to deploy connect straight to Azure AD from the
OSs to devices? OOB experience. They immediately come
under device control, are hardened and
have applications deployed to them.
194

..MS2019..
 The Technical Questions > Monitoring, Management and Automation

 Question Why this matters

⎕ 14.5.12 How With a move away from traditional Active
comfortable are Directory to Azure Active, management
you with Cloud and security of those devices is now
domain-joining delivered exclusively from the Cloud.
your devices?
⎕ 14.5.13 Are you There are two versions of Windows 10 –
aware of the semi-annual and long-term service branch.
release and Microsoft are encouraging users onto the
support cadence semi-annual branch. The support model
for Windows 10? now dictates the September release will
have support for 30 months and the March
release will have support for 18 months.
Consider how you will deal with this rapid
update cycle to the underlying operating
system.
⎕ 14.5.14 How do Automatic deployment of software
you install installation packages is commonplace. The
applications on deployment is done via GPO and/or
devices today? platforms like System Center Configuration
Manager (SCCM). With modern
management, these packages are deployed
from the Cloud. You need to consider the
amount of work involved in changing your
application deployment methodology.
⎕ 14.5.15 What Most of the AV providers have embraced a
anti-virus Cloud world and their solutions play nicely
software do you with the Cloud. Some others do not. The
use today? capabilities of Windows Defender have
also improved a lot over the past few years,
which may negate the need for third party
solutions.

195

..MS2019..
The Technical Questions > Monitoring, Management and Automation

 Question Why this matters

⎕ 14.5.16 Do you This is an important design point and has
envisage implications for the design of your
substantial monitoring and management solution. If
portions of your the majority of your estate will live in the
server Cloud in the near future, it may be
infrastructure appropriate to use Cloud-native technology
remaining on alone, such as Azure Log Analytics. If lots
premises? of on-premises equipment will remain, you
may need to keep System Center or an
equivalent alongside it, at least for now.
⎕ 14.5.17 How There are a lot of things you can automate
much on premises, especially when combined
automation do with virtualization technologies. If you’re
you do today? doing this today, great. You’ll be doing a
whole heap more of it in the Cloud. You
will have the ability to automate things that
you can’t do on premises today.
⎕ 14.5.18 What is PowerShell is the scripting language of
your team’s level choice within the Microsoft Cloud. If your
of skill with team are not already whizzes on it, they
PowerShell? need to become whizzes. Fast. Consider
how you can support your team to upskill
on this technology.
⎕ 14.5.19 How do Alert management is difficult. How do you
you respond to filter the noise of warnings and spot the
alerts today? critical incident that needs to be fixed? This
is one area where the Cloud can assist with
machine learning, spotting the trends and
finding that needle in the haystack.
⎕ 14.5.20 To what If there are things you commonly have to
extent can you do to fix an application that is
take corrective malfunctioning, such as forcing a config
actions in update, can you automate this? Tools such
relation to an as Azure Automation can be programmed
alert? with complex logic about what to do in
different situations, and what to do to fix a
problem.

196

..MS2019..
 The Technical Questions > DevOps and Application Development

14.6 DevOps and Application

Development
Restructuring your organization to take advantage of DevOps is
one of the most challenging aspects of taking on a Digital
Operating Model.
It will have the biggest impact on your organizational structure of
any of the changes you need to make. It will also be the set of
changes which helps you best deliver and execute your Digital
Operating Model.
If every business needs to become a software business, then every
business really needs to build a DevOps culture and team within
their organization.

197

..MS2019..
The Technical Questions > DevOps and Application Development

 Question Why this matters

⎕ 14.6.1 Do you If you don’t currently develop software,
have an internal this can be an advantage as you won’t have
development an existing team to morph.
function?
⎕ 14.6.2 Do you If you don’t currently develop software, the
want an internal question “do you want to” is a very
development important one. Getting into software
function? development is not something to be done
lightly. But it is becoming more necessary
in order to digitally differentiate yourself.
⎕ 14.6.3 Do you Most organizations that do development
have a software today, will have an SDLC in place. You
development might use Team Foundation Server or the
lifecycle Atlassian stack. You will need to make
management much use of this SDLC tool as part of your
(SDLC) tool in DevOps initiative. Make sure you have one
place? that properly supports it.
⎕ 14.6.4 Do you There is no right answer to this question.
have preferences With Microsoft’s Cloud there is no
for a software preference for one language or framework
development over another. .NET is the typical
language and framework used by Microsoft shops, but it
frameworks? doesn’t have to be: Java, Node, Go, Ruby
or any other framework work seamlessly. If
you have no preference, look to the
availability and cost of resources to
determine your best choice.
⎕ 14.6.5 Do you If you do, you may need to fundamentally
have an rethink how they work alongside your
established developers. The role of the operations
operations team team will need to change. Start by
supporting your evaluating their skill sets and their level of
developers? competency in scripting and automation.

198

..MS2019..
 The Technical Questions > DevOps and Application Development

 Question Why this matters

⎕ 14.6.6 Do your It’s vital that development and operations
development and report to the same person. This is the only
operations way to truly embrace DevOps. If they
reporting lines don’t, collapse the teams and put one
converge? leader in charge of both teams.
⎕ 14.6.7 Can you Along with changes to roles, the biggest
sustain rapid change in adopting DevOps is enabling
code releases? rapid change. You need to understand if
your codebase can support this.
⎕ 14.6.8 Do you If you want to create an environment of
have a modular rapid change and your current
architecture? application(s) have a monolithic
architecture, you may want to consider re-
architecting them.
⎕ 14.6.9 Do you Most software defects can be traced back
have centralized to configuration management issues. If you
configuration don’t have centralized configuration
management? management, you need it.
⎕ 14.6.10 How do In a world of DevOps, the traditional
you manage notion of environments fades away. With
environments ephemeral environments, you can have as
today? many or as few environments as you want.
Consider how many environments you
need and how you can support them.
⎕ 14.6.11 Do you In a world of automated deployment, you
utilize third party need to consider how easy it is to bundle
components or and deploy third party libraries. You also
libraries? need to understand any licensing
implications of having more environments
than you might have used previously.
⎕ 14.6.12 Is your If your application is hosted inside another
application based platform, such as SharePoint or Sitecore,
on third party you will need to research how to support
platforms? automated deployment in this model.

199

..MS2019..
The Technical Questions > DevOps and Application Development

 Question Why this matters

⎕ 14.6.13 Have you Containers are the cool kids on the block
investigated right now. There can be benefits from
containers? containerization, but there are also some
drawbacks. Additional platforms will be
required to orchestrate containers.
⎕ 14.6.14 Can you Wherever possible, look to leverage PaaS
leverage platform components as part of your application
services? architecture. This makes deployments
easier and maintenance more
straightforward.
⎕ 14.6.15 What There are many models for building and
delivery platform deploying code today. You might choose
do you want to PaaS, containers, functions or something
use? else. Make the choice carefully as it will
have a profound impact.
⎕ 14.6.16 To what It’s easy to say you are not in any way
extent are you prepared to accept lock-in. But understand
prepared to the ramifications of this statement. Some
accept lock-in? lock-in is sometimes OK if it can reduce
development effort. Likewise, if you assess
the cost of reversing out of an
environment to be less than the cost of
making it work within a different Cloud in
the future it’s an easy choice to make.
⎕ 14.6.17 Do you Most applications can function within the
have availability SLAs offered by Microsoft. This is so long
requirements as the correct design patterns are adhered
beyond that to and you utilize multiple datacenters.
supplied by Some applications simply must never be
Azure? allowed to go down. In this circumstance,
you might want to investigate hosting
across different Cloud providers.

200

..MS2019..
 The Technical Questions > DevOps and Application Development

 Question Why this matters

⎕ 14.6.18 What is One of the prerequisites for DevOps is
your approach to automated testing. It’s the safety net that
testing? allows rapid deployments. If you don’t
currently have unit tests, you need to
investigate the feasibility of introducing
them.
⎕ 14.6.19 Can you Testing code is great. To really rest safe in
automate your the knowledge that changes are not
UI testing? breaking things, you must implement UI
testing.
⎕ 14.6.20 Can you Along with testing the UI, also consider
automate how you can automate the end-to-end
integration testing of your solution in place. You may
testing? need to create testing (mocked) end points
to help with this.

201

..MS2019..
The Technical Questions > DevOps and Application Development

202

..MS2019..
 Funny you should say that

Chapter

15
15Funny you should say that
Laughter gives us distance. It allows us to step back from an event, deal with it
and then move on.
Bob Newhart (Comedian, 1929 – )

F OR many organizations a credible Digital Operating Model

cannot come soon enough. It is the missing piece for
organizations with increasing (and in some cases total)
reliance on Cloud services. They need to develop an
organizational culture that has the business teams and IT team
working in harmony. It will also help those organizations that have
dipped their toes in the Cloud and then thought: What are we
getting into? If you have yet to start your journey in the Cloud
then: What have you been waiting for?
Notice the yellow sticky on the book cover with the three words:
Secure. Compliant. Agile.
 Secure: We tell you all the capabilities to switch on in order
to secure your environment
 Compliant: Ensure your IT environment meets your
governance and policy requirements (e.g. for data
encryption and data residency)
 Agile: This is the crux of the book. It has to do with how
you deliver agility to the business team and control for the
IT team. It is the IT team that has day-to-day responsibility
for this strategic resource.

203

..MS2019..
Funny you should say that

The best person to ask for advice is someone with experience. Find
someone who has climbed the mountain, not just looked at the
pictures, thinking: One day, maybe. There now follow two case-
studies from people who have the experience. They chose to climb
the mountain rather than look at the pictures.
If you have your own story, then we encourage you to share it at
[email protected]

John Kendrick
Cloud Transformation Lead, International Oil Company
Moving to “Cloud” was (and I should say still is) an emotional and
challenging journey. The classic PPT (People Process Tools)
activity goes into overdrive… Do not underestimate the impact,
effort and how long this will take! That said, the benefits are huge.
Classic infrastructure generally takes weeks or months to setup kit
in data centers with many teams involved. It’s expensive and
slow… and often processes used have been in place for many,
many years.
The Cloud proposition is significantly better … but the disruption
it brings is significant too! Creating vanilla infrastructure in the
Cloud takes tens of minutes, but, all the hardening processes, port
opening, active directory setup, everything that exist around it is
from a time when infrastructure took weeks and months to get
setup, and generally geared to have an SLAs (Service Level
Agreements) that is setup to respond within in days (if not weeks);
the SLAs are just not ready, and because of this, nor are the people.
When I embarked on it, we didn’t realize quite how much
disruption we would cause. We spent a long time doing what we
could to automate as much as we could within our new processes.
Automation of process within your own control should be non-
negotiable – anything that involves your new Cloud platform must
be automated as much as possible – remember, this is
Infrastructure as Code, so treat it like that. However, no matter
how fast you go, no matter how much automation you bring
through your new Cloud construct, you can’t get away from the
fact that there are significant changes needed to “external” teams
and processes. Assuming your company follows it, ITIL v3 will

204

..MS2019..
 Funny you should say that

need to be interpreted in a very different light to ensure its

alignment to the speed of Cloud.
Even when you battle through and as you begin to change the
people, process and technology, you encounter the next challenge –
the operating model. There are two elements to this. One, having a
customer focus is really core – you need to create a service people
want to use. Amazon and Microsoft have different technology
offers, but they do have a very common theme of making it
something that the customer wants and desires. The second is
ensuring you are ready to have a service that can keep pace with the
new Cloud offer – and very importantly be able to “resell” it within
your organization. When we started on the Cloud journey
Microsoft were just introducing ARM (moving away from Classic)
and Amazon were churning out pretty significant improvements on
their core products on a daily basis. We had to create an operating
model that could leverage these new features, and to do this we
developed a “product” based model, rather than a “project” based
model to support the Cloud offers. The most significant challenge
in this area was the release schedule – we were releasing
improvements to enterprise Cloud infrastructure every two weeks;
this was across both Amazon and Azure. A lot of people saw this
as a huge risk – but it was actually essential… in fact, if we didn’t
keep pace, there was a higher risk if we stayed on older version of
the various product offers.
Cloud offers amazing capabilities and with it HUGE (and I really
do mean big) positive disruption. This isn’t throwing a pebble into
a pond and watching the ripples move out. This is almost akin to
throwing a boulder so large in that it dislodges all of the water!
Metaphorically the new ways of working are the water in which you
will refill.
It’s a brilliant move. One I fully advocate, and, had I known what I
know now back when I embarked on the journey all those months
ago, I probably would have been more nervous about my chance of
success. That said, I don’t think I would have changed anything
with my approach – “do something, learn, do again, learn again,
etc.”. It was tough, it was hard work, but I was successful, and
Cloud can (and if done properly will) be a positive impact on your
company.

205

..MS2019..
Funny you should say that

Pete Gatt
CEO, Vibrato & Partner, Servian
I’d love to tell a story about RLB in Melbourne, Australia. With a
corporate history stretching back to the Industrial Revolution in
the United Kingdom, Rider Levett Bucknall are market leaders in
costs and project management for major building developments,
including quantity surveying and advisory services.
RLB wanted efficiency in building and maintaining lots of
environments...
Just liked described in chapter 13, the RLB team were operating
with a small infrastructure team who whilst familiar with modern
services within Azure to assist with automating their pipelines,
weren’t ready for production hand over. This is a tale that talks to
the flow of developers and application specialists flowing into the
platform services team, working with a partner and then flowing
back into the application team, allowing Azure to maintain the
platform management with minimal human and manual interaction.
As part of its core business, Rider Levett Bucknall (RLB) enables
surveyors and project managers to operate efficiently on landmark
building projects such as the Sydney Opera House and the 2012
London Olympics. This requires access to a scalable and robust
data platform which is easily consumable.
When we got to RLB, we found that they had been working with
US-based data specialists to develop a data platform completely
customized to support its business workflows. This data platform
was built and deployed in Microsoft Azure, and required a highly
scalable, geo-redundant architecture to ensure that the custom
solution could be accessed reliably, could scale rapidly to meet
increasing demand, and could be quickly deployed into multiple
global regions. Whilst the application was highly sophisticated, the
platform and automation of environment creation and
management wasn’t in place. The teams had amazing developers
and at the time, only a small team who were new to Azure and new
to looking after these workloads. But this meant that building
environments and deploying to environments was looking like days
or weeks at a time for them. We (Vibrato, now Servian) thought,
there’s got to be a way we can continue to ensure they keep it
simple and don’t require a massive amount of engineering and
206

..MS2019..
 Funny you should say that

continued engineer to look after this platform. So, we established a

virtual platform services team in partnership with RLB and built on
a foundation of Microsoft Azure Container Services, leveraging
Kubernetes for container orchestration. There’s more to the story
and process, but the solution has enabled RLB to deploy changes
from code commit into production across regions in minutes,
whilst providing a unified environment for internal development
teams.
We must scale this beast... globally...
The solution that we architected needed to adhere to certain critical
business requirements, including database scalability and
redundancy, automated deployment to Dev/Test environments
and integrated approval gateways for production deployments. The
development teams did not want to have to keep spinning things
up differently in each region and didn’t want to have to redeploy
things manually each time... nor did we. We realized that if we’re
going to scale it, we’re going to scale it in tiers. The solution was
architected and delivered in three discrete components – an
application tier built on Azure Container Services, a data tier
leveraging Microsoft Azure SQL and an automated CI/CD stack
founded on Microsoft DevOps. Working with RLB’s development
team which drove the internal adoption and development of
application containers, Vibrato architected a fully automated
solution which was redundant across multiple Azure regions with
the capability to easily expand to any region globally.
The container-driven application tier was architected using
Kubernetes for container orchestration, allowing RLB to define its
entire application stack as source-controlled templates and maintain
a unified deployment pattern across every environment and every
Azure region.
So, about this CI/CD and DevOps Stuff...
To ensure that we minimized the requirement for people managed
servers or services, we needed to ensure that every phase of the
solution was delivered with strict adherence to the principles of
Continuous Delivery, leveraging the native capabilities of Microsoft
Azure DevOps (AZDO). Using Builds triggered from code
committed to both GitHub and AZDO Git repositories, RLB was
able to build and store custom application containers in Azure
207

..MS2019..
Funny you should say that

Container Registry instances, then launching Releases with the

same image across Development, Staging and Production
environments. We had given the developers a simple way to
leverage event driven builds through their SDLC seamlessly.
Kubernetes was used heavily for container orchestration, ensuring
not only that the same image could be dynamically configured for
each deployment environment, but that each service was self-
healing and could be updated during business hours without end-
user impact. Kubernetes was also used in combination with
Microsoft Azure Traffic Manager to dynamically route users to the
region with the lowest latency, whilst also providing secure
communications and dynamic edge routing.
The native capability of AZDO to integrate with open source
platforms such as Docker and Kubernetes ensured that RLB was
able to build upon a single unified platform for Continuous
Integration and Continuous Deployment, whilst maintaining the
flexibility to deploy from a mixed pool of Windows and Linux
agents.
Remember the part where they didn’t want to have to maintain
environments, this well and truly solved that and the Platform
Services capability was handed to development teams to continue
using not just for this service, but for the other services to come.
Deploying in minutes to production...
RLB’s business and development teams were extremely happy with
the solution, in particular the cross-region robustness and
redundancy and the fully automated AZDO Build and Release
pipelines. The abstracted, templated nature of the solution means
that RLB’s development teams were freed from ongoing
infrastructure maintenance and can focus entirely on supporting
the application and supporting the business by driving further
innovation.
The adoption of container services and adherence to DevOps
principles has enabled RLB to deploy application updates from
code into production in mere minutes, whilst using the native
capabilities of AZDO to maintain tight control over the end-to-end
process.
And thus, RLB is now able to drive innovation and expansion,
confident in the underlying technical foundations, giving the power
208

..MS2019..
 Funny you should say that

back to the developer and letting the Azure platform take care of
the rest.
So, what did we learn?...
 The harmonic operating model does not need to be fixed in
place once established. In fact, Servian has since found that
it’s often better to establish and destroy platform service
squads when paying off a NTOP TAX (new technology or
process tax) rather than keep a team in place for good. In
drives the ability for teams to not get fat and lazy and
ensure they hand back solid IP, automation and process
back to the development squads.
 It’s possible to get your deployment down from weeks to
minutes … even to prod when using Azure DevOps in a
highly integrated fashion.
 Kubernetes for the win! Allowing Azure to manage the
clusters without breaking the back is far cheaper than
building out an operations team.
 Models that hand the control back to the developer are
good models. Developers are the Rockstars, we’re just the
Roadies in the background ensuring they put on an
amazing performance for their fans.

209

..MS2019..
Funny you should say that

210

..MS2019..
 Final Word

Chapter

16
16Final Word
A conclusion is the place where you got tired of thinking.
Albert Bloch (American Artist, 1882 – 1961)

F OR the business audience, we live in interesting times. This

book was inspired by conversations with Microsoft and
Microsoft customers and partners who had reached a point
of no return along their Cloud journey. The evidence was for
them compelling: Cloud is the future; it is where innovation lives; it
rocks!
This in turn raised important questions: What does that mean for
the existing IT estate? How should we deal with the rapidly
increasing adoption of Cloud with a likely end state of being ‘all in’
on the Cloud? Some businesses are already committed to being ‘all
in’. Others are starting out, and for them unchartered waters lie
ahead.
Part 1 of this book dealt with the conversations that surface as a
business grows its dependency on the Cloud. It presents the
argument for a Digital Operating Model. We broke that term down
into the language of the business team and IT team. The Digital
Operating Model is a discussion between the business teams and
IT team wanting a ‘let’s make this happen’ approach instead of
‘locking horns’.
Perhaps we will look back one day and ask: What was all the fuss
about? For now, there is an urgent need for practicality, and we
hope this book offers that practical guidance.

211

..MS2019..
Final Word

F OR the technical readers; we’ve covered a phenomenal

amount of ground together in Part 2 of the book.
We have delivered a soup-to-nuts journey from strategizing,
through buying, managing, securing, monitoring and developing
your new Digital Operating Model.
If you want one thing to remember from this whole discussion
around agility and control, it’s automation and code. Full, extensive
and comprehensive automation is the secret sauce. It’s what will
allow you to discover that enigma that has escaped us for so long.
Your journey has just begun though. Getting from the theory in
this book to the practice inside your organization is no mean feat.
Depending on the size of your business, it might take anywhere
from months to years. You may have decades of legacy you need to
remodel and adapt. The journey will be fraught with challenges,
both political and practical. You may have to crack a few eggs. You
may have to bring a few new people in and let a few people go.
This is something you will almost certainly need help with. Reach
out. Speak to others who have done it before. Seek guidance,
counsel and help. Get in experts who have helped other
organizations and can assist in avoiding the potholes and blind
alleys.
Please do not be put off though. If you can pull through and
deliver your new Digital Operating Model, you will be setting your
organization up to succeed not only now, but for the next decade.
You will deliver on what was only a dream up to now, combining
agility and control. This will put technology at the core of what
your organization does. It will transform you into a truly digital
business.
We finish up with just two words.

Good luck!

212

..MS2019..
 Glossary

Glossary

Agile
Agile is a methodology for modeling and documenting software
systems based on best practices. It is a collection of values and
principles, that can be applied on an (agile) software development
project. This methodology is more flexible than traditional
modeling methods, making it a better fit in a fast-changing
environment. See Chapter 5.

agile
Small ‘a’ is used to describe an organization’s intention to establish
an agile culture for the collaboration of teams working to achieve a
common goal.

Artificial Intelligence (AI)

AI is hot technology. It is the development of computer systems
able to perform tasks normally requiring human intelligence, such
as visual perception, speech recognition, decision-making. The
Turing Test is a method that is a reference for determining whether
or not a computer is capable of thinking like a human being.
AI is highly topical today and subject of critical debate about its
societal impact and ethics. Here we reference Microsoft’s
submitted written evidence to the House of Lords Select
Committee on AI.
The human-centered approach to AI that Microsoft envisions can
only be realized if relevant stakeholders from industry, government,
civil society and the research community collaborate on the
development of shared principles to shape the use of AI
technologies.
Microsoft’s CEO, Satya Nadella, shared some initial thoughts on
what these may be in order to start this dialogue. We believe that
AI should:
1. Be designed to assist humanity;
2. Be transparent;

213

..MS2019..
Glossary

3. Maximize efficiencies without destroying the dignity of

people;
4. Be designed for privacy;
5. Have algorithmic accountability so that humans can undo
unintended harm;
6. Guard against bias.
Full report (1600 pages) at https://www.parliament.uk/documents/lords-
committees/Artificial-Intelligence/AI-Written-Evidence-Volume.pdf

BAU
Business As Usual. In a world that is seemingly under assault from
constant change, BAU is getting harder to define as a steady state
of affairs. Perhaps BAU is really all about the acceptance of change.

Big Data
Big Data is characterized by extremely large data sets that may be
analyzed computationally to reveal patterns, trends, and
associations.
It is well known that we are in the midst of a data explosion and
Big Data is about how that data will be processed to extract its
value, some say providing insights that will bring change to the way
we run business and society. It also needs people who can interpret
the data and therein lies a current skills shortage.

Business Model
Not to be confused with Business Operating Model. The Business
Model describes the way the business generates value and the
raison d’être of their existence. Many businesses fail when their
business model becomes out of date and others steal their
customers with a superior (in value) business model. There are
examples in the book.

Business Operating Model

This describes the way a business structures its core processes and
will vary between industry sector and businesses operating in the
same industry sector. It has the potential to deliver competitive
differentiation and in a modern business is reliant on technology.
214

..MS2019..
 Glossary

DevOps
DevOps is the union of people, process, and products to enable
continuous delivery of value to our end users. The contraction of
“Dev” and “Ops” refers to replacing siloed Development and
Operations to create multidisciplinary teams that now work
together with shared and efficient practices and tools. Essential
DevOps practices include agile planning, continuous integration,
continuous delivery, and monitoring of applications.

Source: https://docs.microsoft.com/en-us/azure/devops/what-is-devops

Digital Operating Model

The Digital Operating Model is where everyone comes together,
Business teams, Developers, IT Operations. It’s where discussions
about ‘digital at the core’ of delivering business outcomes get
turned into actionable plans. Those actionable plans become the
end to end blueprint of what you are and how you operate as a
digital business with change as the only constant.

Infrastructure as a Service (IaaS)

The capability provided to the consumer is to provision processing,
storage, networks, and other fundamental computing resources
where the consumer is able to deploy and run arbitrary software,
which can include operating systems and applications. The
215

..MS2019..
Glossary

consumer does not manage or control the underlying cloud

infrastructure but has control over operating systems, storage, and
deployed applications; and possibly limited control of select
networking components (e.g., host firewalls).
Source: https://nvlpubs.nist.gov/

IT Operating Model (ITOM)

Depending on the size of your organization and reliance on IT the
ITOM may be documented in a few pages or in volumes of
documents. The ITOM aligns the business strategy with the day-
to-day operations of IT. It is known to few people at the heart of
IT operations. There are plenty of disagreements about the
definition of an ITOM. One thing is for sure, if you ask 10
employees what the ITOM is, you will probably get ten blank faces.
Ask them what they use the Internet for, and you will likely have a
long conversation.

Machine Learning (ML)

This book makes many references to Artificial Intelligence (AI).
Sometimes there is confusion about the difference between AI and
ML. ML is where a machine learns to perform a specific task
without using explicit instructions, relying on patterns and
inference instead. It is seen as a subset of Artificial Intelligence.

Own and Operate

Many moons ago the option to acquire IT was to own and operate
and staff an IT department. There remains much investment in IT
hardware and software on the books of businesses being written
down. The Cloud provides an economic alternative to the capital
intensive own and operate model.

Platform as a Service (PaaS)

The capability provided to the consumer is to deploy onto the
cloud infrastructure consumer-created or acquired applications
created using programming languages, libraries, services, and tools
supported by the provider. The consumer does not manage or
control the underlying cloud infrastructure including network,
216

..MS2019..
 Glossary

servers, operating systems, or storage, but has control over the

deployed applications and possibly configuration settings for the
application-hosting environment.
Source: https://nvlpubs.nist.gov/

Prince 2(TM)
PRINCE2 (an acronym for PRojects IN Controlled
Environments) is a de facto process-based method for effective
project management. Used extensively by the UK Government,
PRINCE2 is also widely recognized and used in the private sector,
both in the UK and internationally. The PRINCE2 method is in
the public domain and offers non-proprietorial best practice
guidance on project management.
Key features of PRINCE2:
 Focus on business justification
 Defined organization structure for the project management
team
 Product-based planning approach
 Emphasis on dividing the project into manageable and
controllable stages
 Flexibility that can be applied at a level appropriate to the
project.
Source: https://www.prince2.com/uk/what-is-prince2

Robotic Process Automation (RPA)

In traditional workflow automation tools, a software developer
produces a list of actions to automate a task and interface to the
back-end system using internal application programming interfaces
(APIs) or dedicated scripting language. In contrast, RPA systems
develop the action list by watching the user perform that task in the
application's graphical user interface (GUI), and then perform the
automation by repeating those tasks directly in the GUI. This can
lower the barrier to use of automation in products that might not
otherwise feature APIs for this purpose.
Source: Wikipedia

217

..MS2019..
Glossary

Software as a Service (SaaS)

The capability provided to the consumer is to use the provider’s
applications running on a cloud infrastructure. The applications are
accessible from various client devices through either a thin client
interface, such as a web browser (e.g. web-based email), or a
program interface. The consumer does not manage or control the
underlying cloud infrastructure including network, servers,
operating systems, storage, or even individual application
capabilities, with the possible exception of limited user specific
application configuration settings.
Source: https://nvlpubs.nist.gov/

Waterfall
The waterfall model is a relatively linear sequential design approach
for certain areas of engineering design. In software development, it tends to
be among the less iterative and flexible approaches, as progress
flows in largely one direction ("downwards" like a waterfall) through
the phases of conception, initiation, analysis, design, construction,
testing, deployment and maintenance.
The waterfall development model originated in the manufacturing
and construction industries; where the highly structured physical
environments meant that design changes became prohibitively
expensive much sooner in the development process. When first
adopted for software development, there were no recognized
alternatives for knowledge-based creative work.
Source: Wikipedia.

218

..MS2019..
 List of Abbreviations

List of Abbreviations
ACL Access Control List
ADFS Active Directory Federation Services
ASC Azure Security Center
AWS Amazon Web Services

CDO Chief Digital/Data Officer

CEO Chief Executive Officer
CI/CD Continuous Integration / Continuous Deployment
CIO Chief Information Office
CISO Chief Information Security Officer
COM Cloud Operating Model
COTS Commercial Off The Shelf
CSP Cloud Solution Provider
CTO Chief Technology Officer

DDoS Distributed Denial of Service

DSC Desired State Configuration

EA Enterprise Agreement

FOM Future Operating Model

GCP Google Cloud Platform

GPO Group Policy Object
HANA SAP Brand Name
HOD Head of Delivery

219

..MS2019..
List of Abbreviations

HPC High Performance Computing

IdP Identity Provider

ITOM Information Technology Operating Model
JML Joiners, Movers, Leavers
JSON JavaScript Object Notation
KPIs Key Performance Indicators

LSP Licensing Service Provider

MTP Microsoft Threat Protection
NIST National Institute of Standards and Technology
NoSQL No SQL
NPS Net Promoter Score

OOB Out of the Box

PCI DSS Payment Card Industry Data Security Standard

REST Representational state transfer

ROI Return on Investment
RRP Recommended Retail Price

SAP S4 SAP Brand Name

SDLC Software Development Lifecycle Management
SIAM Service Integration and Management
SKU Stock Keeping Unit
SLA Service Level Agreement
SOAP Structured Object Access Protocol

220

..MS2019..
 List of Abbreviations

SSO Single Sign-on

TFS Team Foundation Server

TOGAF The Open Group Architecture Framework

UAT User Acceptance Testing

ULA User Level Agreement

VLA Value Level Agreement

VM Virtual Machine
VPN Virtual Private Network

221

..MS2019..
List of Abbreviations

222

..MS2019..
 List of Figures

List of Figures
Figure 1 - Disruption by industry ......................................................... 4 
Figure 2 - The 4th industrial revolution ............................................. 18 
Figure 3 - Transformation agenda roles ............................................. 21 
Figure 4 - Azure Service Map .............................................................. 30 
Figure 5 - Agile funding models .......................................................... 40 
Figure 6 - SAFe® Portfolio Configuration ©
scaledagileframework.com ................................................................... 44 
Figure 7 - Guiding principles for IT Strategy .................................... 84 
Figure 8 – Bimodal IT© Gartner Inc ................................................. 89 
Figure 9 - Cloud Maturity Model ........................................................ 90 
Figure 10 - ITIL Service Map ............................................................ 105 
Figure 11 - Traditional SIAM Service Model ©
www.itforbusiness.org ........................................................................ 106 
Figure 12 - Portfolio-based model .................................................... 108 
Figure 13 - Provisioning paths .......................................................... 111 
Figure 14 - Code libraries ................................................................... 113 
Figure 15 - Traditional support model ............................................. 117 
Figure 16 - SaaS support model ........................................................ 117 
Figure 17 - Traditional application support model © Servian ...... 118 
Figure 18 - Next generation applications support © Servian ....... 119 
Figure 19 - Cloud adoption and governance teams ....................... 119 
Figure 20 - Azure security model ...................................................... 128 
Figure 21 - Azure role-based access hierarchy ................................ 129 
Figure 22 - Trust and control model ................................................ 130 
Figure 23 - Azure ExpressRoute conceptual diagram ................... 131 
Figure 24 - Monitoring map .............................................................. 141 
Figure 25 - The DevOps problem statement © Servian ............... 149 
Figure 26 - The first, second and third way of communications . 159 
Figure 27 - A typical DevOps pipeline............................................. 160 
Figure 28 - Azure Application Insights .......... Error! Bookmark not
defined. 

223

..MS2019..
Getting Involved

Getting Involved

The Smart Questions community

There may be questions that we should have asked but didn’t. Or
specific questions which may be relevant to your situation, but not
everyone in general. Go to the website for the book and post the
questions. You never know, they may make it into the next edition
of the book. That is a key part of the Smart Questions Philosophy.

Send us your feedback

We love feedback. We prefer great reviews, but we’ll accept
anything that helps take the ideas further. We welcome your
comments on this book.
We’d prefer email, as it’s easy to answer and saves trees. If the
ideas worked for you, we’d love to hear your success stories.
Maybe we could turn them into ‘Talking Heads’-style video or
audio interviews on our website, so others can learn from you.
That’s one of the reasons why we wrote this book. So talk to us.
[email protected]

Got a book you need to write?

Maybe you are a domain expert with knowledge locked up inside
you. You’d love to share it and there are people out there
desperate for your insights. But you don’t think you are an author
and don’t know where to start. Making it easy for you to write a
book is part of the Smart Questions Philosophy.
Let us know about your book idea, and let’s see if we can help you
get your name in print.
[email protected]

224

..MS2019..
 Notes pages

Notes pages
We hope that this book has inspired you and that you have already
scribbled your thoughts all over it. However, if you have ideas that
need a little more space then please use these notes pages.

225

..MS2019..
Notes pages

226

..MS2019..
..MS2019..