Storage

Virtualization
FOR

DUMmIES
‰

HITACHI DATA SYSTEMS EDITION

by Adrian De Luca and Mandar Bhide

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

01_597705-fm.indd i 9/30/09 10:44:17 AM

 Table of Contents
Introduction ....................................................... 1
What’s in This Book ...................................................................2
Icons Used in This Book ............................................................3

Chapter 1: Storage Virtualization: Getting Started . . . . . . . . . 5

Defining Virtualization ...............................................................5
Looking for What Characterizes an Ideal System ..................7
Storage Virtualization: A Stroll Down
Memory Lane...........................................................................7
Drivers for Storage Virtualization ..........................................10
Out of storage again! ......................................................11
Reducing cost .................................................................12
Managing service levels ................................................12
Adding agility and flexibility .........................................13
Repurposing and adding value to
existing assets .............................................................14
Going “green” ..................................................................15

Chapter 2: Types of Storage Virtualization . . . . . . . . . . . . . . . 17

Storage Virtualization in Three Layers of
Infrastructure ........................................................................19
In the server ....................................................................19
In the storage network ..................................................22
In the storage controller ...............................................26
Choosing the Right Solution ...................................................29
What are your objectives? ............................................29
Where are you today? ....................................................29
Price versus cost ............................................................30
Matching growth to scalability .....................................30
Can you afford an outage? ............................................30

Chapter 3: The Economics of Storage

Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
How Do You Measure the Economic
Benefits? .................................................................................32
So Where Are the Savings in Storage? ...................................34
Where Do You Start? ................................................................35
Hitachi’s Storage Economics Model.......................................36

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

01_597705-fm.indd v 9/30/09 10:44:18 AM

 vi Storage Virtualization For Dummies

Chapter 4: Making Storage Virtualization Real. . . . . . . . . . . 39

Key Concepts in Storage Virtualization .................................40
Planning for the Virtualized Storage Environment ..............41
Implementing Storage Virtualization .....................................42
Partitioning of storage hardware .................................43
Internal and external storage configurations .............43
Migrating Data to a Virtualized Storage
Environment ..........................................................................45
Classifying the data types .............................................46
Creating storage tiers ....................................................46
Uninterrupted data migration ......................................47

Chapter 5: Getting the Best from Storage

Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Optimizing Performance and Improving
Storage Utilization ................................................................49
Understanding thin, or dynamic,
provisioning.................................................................50
Creating a dynamic provisioning pool ........................50
Benefits of dynamic provisioning ................................51
Management Challenges in the Virtualized
Storage Environment ............................................................53
Dealing with the challenges ..........................................55
Managing performance and capacity ..........................56

Chapter 6: Ten Best Practices for Deploying Storage

Virtualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
#1: Choose the Right Storage for
the Host’s Mapping ...............................................................57
#2: Assign External Storage Configurations ..........................58
#3: Format the Virtual Disk .....................................................58
#4: Choose the Right Disk Attributes for
Data Migration .......................................................................59
#5: Migrate the Data .................................................................59
#6: Classify the Data .................................................................60
#7: Understand the Workload or I/O Profile .........................61
#8: Design the Dynamic Provisioning
Storage Pool...........................................................................62
#9: Create a Dynamic Provisioning Pool ...............................63
#10: Plan and Design a Virtualization Project ......................63

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

01_597705-fm.indd vi 9/30/09 10:44:18 AM

 Foreword
W hen I was a young man, fresh out of college, I could
throw all my worldly possessions into the back of
my Chevrolet and go wherever I needed to be. Since then, I
got married, had a family, bought a house, bought furniture,
appliances, and tons of other stuff to support my family’s
lifestyle, and now it would take several truckloads to move me
with all my stuff. I have lived in the same community for the past
20 years and have built a life here. It would be difficult for me
to move if my work required it. However, with the availability of
the Internet and broadband, my workplace has been virtualized
and I can do my work without disrupting my life and moving all
my stuff.

Just like my life, businesses are burdened with tons of stuff,

mainly data, that they must store, preserve, protect, and access.
That burden is nearly doubling every year and is beginning to
strangle businesses, limiting their ability to do their work.

Data center managers are handling aging applications that are

running on aging infrastructures and facilities that were built
when electricity was readily available and carbon footprints were
not a concern. They spend more and more of their data center
budget on maintaining old systems instead of investing in new
systems that are more efficient and environmentally friendly. The
reason it is so difficult to modernize IT is the burden of data that
must be carried forward with the modernization of applications
and infrastructures. Ten years ago, before the turn of the
millennium, changes could be made by suspending operations
during off hours while data was migrated or transitioned to a
new storage platform. Storage frames were directly attached to
application servers and capacities were small enough that they
could be moved within a day.

Since then, the explosion of data, globalization, and online

access requires applications to be available around the clock,
and suspension of operations must be kept to an absolute
minimum. The capacity of storage frames has increased
a hundredfold and many more applications now share
access to these large storage frames, making it difficult to
coordinate downtime for shared storage frames. This makes

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

01_597705-fm.indd vii 9/30/09 10:44:18 AM

 viii Storage Virtualization For Dummies

it nearly impossible to migrate applications or modernize the

infrastructure without a major outage.

In my case, my ability to work was greatly simplified through

virtualization of the workplace. I can do my work anywhere there
is broadband access.

In the business world, there is a similar need to virtualize

storage in order to simplify operations and make it easy to
modernize IT. Only through virtualization of storage can data
centers hope to have the agility to modernize IT without major
disruptions to their business.

To many people, the subject of storage virtualization is very

complex. The purpose of this book is to simplify this complexity,
remove the misconceptions and hype around storage
virtualization, and leave you with a basic understanding of what
storage virtualization is and what it can do for you.

Hubert Yoshida
Vice President and Chief Technology Officer
Hitachi Data Systems

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

01_597705-fm.indd viii 9/30/09 10:44:18 AM

 Introduction
A lot has been written, taught, and even debated about
virtualization in the information technology industry over
the past few years, especially in the server world where it is best
known. Vendors like VMware and Microsoft actively promote
how their virtualization software provides efficient resource
utilization, dynamic capabilities and ease of management, while
the hardware vendors fiercely contend to be the company that
has the lowest power consumption, better interoperability, and
ease of deployment and support.

Vendor “marchitecture” aside, all organizations are looking to

improve efficiency, drive up productivity, and maximize every
dollar of their IT investments. Virtualization has become an
effective way to do more with less, especially in these tough
economic times.

However, for all the hype generated over virtualization, there

remains a considerable number of misconceptions and general
confusion. Many people think virtualization is some new whiz-
bang technology when, in fact, the concept has been around a
long time, with its origins dating back to the mighty mainframe
era. The philosophy of aggregating physical components in a
coordinated way to increase systems utilization and efficiency is
very mature and has been successfully applied in many areas of
IT beyond servers, such as networking and applications.

So what about storage virtualization? Can the same benefits

be realized when virtualization is applied to the storage area
network (SAN)? The fact is that virtualization in the storage
environment shares many of the same benefits of server
virtualization and, most importantly, is highly complementary,
extending agility and cost savings to the data assets.

As a reader of this book, you are most likely looking to

understand what virtualization is in the storage world, what
benefits you are likely to realize, and how best to deploy
virtualization in your environment to get the most out of it.
Storage Virtualization For Dummies, Hitachi Data Systems Edition,
explains the differences in implementation among various
vendors and, more specifically, how it is applied by Hitachi.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

02_597705-intro.indd 1 9/25/09 11:51:15 AM

 2 Storage Virtualization For Dummies

It also tries to dispel the myths to bring an objective look at how

this technology is best applied.

Much like other forms of virtualization, the path is not always

simple. Objectives need to be clearly understood and agreed
at the outset, and a sobering understanding of where you are
today needs to be discovered. In addition, as with any IT project,
preparation and planning needs to be invested in the design, and
the solution correctly implemented, if you are to realize the true
benefits of virtualization.

The good news is that this journey does not need to be walked
alone. A lot of experience has been developed over the five or so
years during which this technology has made its breakthrough
into storage. Many organizations have transitioned their
dysfunctional legacy storage environments into highly flexible
and agile infrastructures that more closely align and adapt
to their business needs. Vendors have the technology, best
practices, and skills to assist you, and resources such as this
book help put you on the right track to succeed.

What’s in This Book

This book comprises six chapters.

Chapter 1 sets the framework of storage virtualization, giving

you a summary of what it is and how you apply it, as well as
an outline of how it came about. This chapter also analyzes
the drivers of storage virtualization (why you’ll even think of
applying it).

In Chapter 2 we talk about the types of storage virtualization

you can choose from and how to select the right type for your
situation. Chapter 3 looks at the economics of storage, including
an outline of Hitachi’s Storage Economics model and a case
study.

Chapter 4 delves into more detail on planning and implementing

storage virtualization according to your own objectives.
This chapter also examines how to migrate your data to the
virtualized storage environment. And in Chapter 5 we explain
how to optimize and manage the system’s performance.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

02_597705-intro.indd 2 9/25/09 11:51:15 AM

 Introduction 3
Last but not least, no For Dummies book would be complete
without a chapter of tens. So Chapter 6 quite succinctly
explains the ten best practices for physical configurations, data
migrations, and thin, or dynamic, provisioning, and tips for
designing your first storage virtualization project.

Icons Used in This Book

Throughout this book we use a series of icons in the margins
that help flag special information. Here’s what to look for.

When we tell you something about storage virtualization that

bears remembering, we mark it with the Remember icon.
Descriptions that appear beside this icon are worth storing away
because they help build your understanding.

This flags a shortcut or easy way to do something.

This icon indicates the information is really for the propeller

spinners; you don’t need to know all about how this works but
you may find it interesting.

This icon indicates a resource on the Internet you can look at to

get further information.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

02_597705-intro.indd 3 9/25/09 11:51:15 AM

 4 Storage Virtualization For Dummies

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

02_597705-intro.indd 4 9/25/09 11:51:16 AM

 Chapter 1

Storage Virtualization:
Getting Started
In This Chapter
✓ Understanding virtualization and how it applies to storage
✓ Identifying what makes an ideal system
✓ Familiarizing yourself with the evolution of storage virtualization
✓ Finding out what drives the need for storage virtualization

V irtualization is one of the biggest buzzwords of the

technology industry right now. But what is virtualization?
More importantly, what can virtualization do for you and your
company?

In this chapter we look at just what virtualization is and how

the concept applies in a practical sense to computer storage
systems. We also take you on a stroll down memory lane to
understand how storage virtualization came about and some
of the business considerations that create the need for storage
virtualization.

Defining Virtualization
If you look on the Internet, you’ll find any number of definitions
for virtualization as it relates to IT.

A general definition of virtualization is:

The act of integrating one or more (back-end) services or functions

with additional (front-end) functionality for the purpose of
providing useful abstractions.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 5 9/25/09 11:51:44 AM

 6 Storage Virtualization For Dummies

Abstracting the physical components of computing resources

optimizes the way in which other systems, applications, or end
users interact with them. This definition summarizes two very
important characteristics of virtualization: It should reduce the
complexity when compared with managing devices discretely, as
well as add greater capability to improve services — a bit like a
one-plus-one-equals-three equation.

When described in terms of storage, the Storage Networking

Industry Association (SNIA), the independent body of the
storage industry, has a more specific definition (two of them,
in fact) for storage virtualization, from its technical tutorial on
virtualization:

1. The act of abstracting, hiding, or isolating the internal

functions of a storage (sub)system or service from
applications, computer servers, or general network
resources for the purposes of enabling application and
network independent management of storage or data.
2. The application of virtualization to storage services or
devices for the purpose of aggregating, hiding complexity,
or adding new capabilities to lower-level storage resources.

Put simply, storage virtualization aggregates storage

components, such as disks, controllers, and storage networks,
in a coordinated way to share them more efficiently among
the applications it serves. Pooling these resources in a logical
way helps remove the physical barriers that would otherwise
exist and maximize the full potential of these resources. Hiding
the complexity associated with all the components required to
deliver storage helps simplify the operations of managing the
environment and, more importantly, provides greater flexibility
to meet the needs of applications.

Think of virtualization in the context of using a telephone.

When you pick up the handset or initiate a call on your mobile
phone, as the caller you are completely unaware of how you are
connected to the other party. The complexity of establishing
communication is completely invisible; you are oblivious as to
the number of hops across the telephone network or the type
of communication involved (such as satellite or underground
cable), or even if there is a failure in the network. This is
virtualization technology at work, using available resources as
efficiently and intelligently as possible to deliver a service.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 6 9/25/09 11:51:44 AM

 Chapter 1: Storage Virtualization: Getting Started 7

Looking for What Characterizes

an Ideal System
Many vendors offer storage virtualization solutions, each
implemented in a different way or offered in different packaging.
Therefore, it is important to understand the characteristics of an
ideal storage virtualization solution.

A good storage virtualization solution should:

✓ Enhance the storage resources it is virtualizing through the

aggregation of services to increase the return of existing
assets.
✓ Not add another level of complexity in configuration and
management.
✓ Improve performance rather than act as a bottleneck in
order for it to be scalable. Scalability is the capability of a
system to maintain performance linearly as new resources
(typically hardware) are added.
✓ Provide secure multi-tenancy so that users and data can
share virtual resources without exposure to other users’
bad behavior or mistakes.
✓ Not be proprietary, but virtualize other vendor storage in
the same way as its own storage to make the management
seamless.

Storage Virtualization: A Stroll

Down Memory Lane
With the introduction of commercial computers back in the
1950s, like IBM’s mainframes, the magnetic disk storage used
to maintain information was directly attached via a cable, or a
bus, to the computer’s central processing unit (CPU) and volatile
random access memory (RAM). This method of attachment,
commonly referred to as direct attached storage (DAS), provided
a modest amount of non-volatile storage to a single computing
system.

Over the course of the next few decades, standard protocols (the
language used to communicate between devices) emerged, such
as the small computer system interface (SCSI), making it easy to

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 7 9/25/09 11:51:44 AM

 8 Storage Virtualization For Dummies

connect different vendor storage to the systems. This method

of connectivity extended beyond traditional magnetic storage
to devices such as CD-ROMs, tape drives and autoloaders, and
JBOD (just a bunch of disks). Although the different types of
storage flourished and fault-tolerant designs provided greater
reliability, their connectivity was still confined to single servers
or workstations, limiting the utilization of the media.

The first wave of storage virtualization appeared in the cache

controller arrays. Large numbers of magnetic disks were pooled
together and arranged in a way to provide fault tolerance to
protect against individual disk failures. This system became
known as RAID (redundant array of independent disks). Utilizing
a common pool of cache memory, applications could work
with a logical image of a data block rather than working with
the actual data block as it spun around on a disk platter. This
improved performance by masking the seek and rotation delays
of a mechanical disk. It also enabled mainframes to use lower
cost disks.

In the early 1980s, in the midst of the computer network

revolution that allowed servers to talk to one another and
exchange information over standard mediums like ethernet,
vendors such as Novell and Sun Microsystems figured that, if
servers can share data, then why not share their storage as
well? By harnessing the power of the network, network attached
storage (or NAS as it is commonly known today), allowed
servers to treat storage in another room, or potentially another
city, as though it was right there, physically attached to the
server. This offered unprecedented levels of flexibility to store
data centrally but make it available widely. It also allowed
multiple users to write and read to the storage at the same time,
providing universal access to the same data sets and increasing
collaboration between users. As this common way to access
data became popular, other operating system vendors started
offering this capability in their systems, and companies like
Network Appliance, EMC, and Hitachi Data Systems packaged the
functionality into appliances for easy deployment and increased
scalability.

The next advancement in storage virtualization came with the

introduction of storage area networks (SANs) in the mid to late
1990s, when the concept of storage resource aggregation made it
out of the box.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 8 9/30/09 10:46:05 AM

 Chapter 1: Storage Virtualization: Getting Started 9

What is a storage area network?

A storage area network (SAN) is an Traditionally, SANs have used optical
architecture to connect detached connectivity called fibre channel (FC)
computer storage devices, such as due to its gigabit transfer speeds and
disk arrays, tape libraries, and optical ability to work over long distances.
jukeboxes, to servers in a way that the However, more recently ethernet-
devices appear as local resources. based storage networks have become
SANs deliver storage to servers at a popular due to the ubiquity (and there-
block level, and feature mapping and fore lower cost) of the technology in
security capabilities to ensure only the TCP/IP world (transmission control
one server can access the allocated protocol/Internet protocol) and the
storage at any particular time. adoption of the SCSI protocol, called
iSCSI (Internet small computer system
The protocol, or language, used
interface) to the medium.
to communicate between storage
devices and servers is SCSI (small
computer system interface).

SANs helped organizations consolidate their storage assets

to improve capacity utilization by sharing their storage
resource effectively. This simplified management by using
common software tools, and enabled replication of critical
information over long distances to provide greater levels
of protection against data corruption and disaster events.
Many larger organizations, typically financial institutions
and telecommunication providers, were among the first to
implement SANs.

Even though these improvements helped organizations extract

better value from their storage investments, a number of
problems still remained. Islands of SANs were created due to
poor interoperability between different vendors and devices,
utilization levels were still relatively low due to traditional
storage allocation practices being used, and the lack of mobility
or replication between different storage vendors meant
organizations were still hamstrung.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 9 9/25/09 11:51:45 AM

 10 Storage Virtualization For Dummies

What is network attached storage?

Network attached storage (NAS) has NAS storage typically connects to
traditionally been a single device that an organization’s TCP/IP network via
contains disk storage and computer ethernet (the same used by servers)
components (CPU, memory, network and uses specific languages, or
ports) and whose sole purpose is to protocols, like NFS (network file
deliver storage to servers. NAS is system) or CIFS (common Internet file
sometimes referred to as a ‘file server’ system), based on the server message
due to its ability to share a common block protocol, to exchange infor-
pool of files with multiple servers. mation between servers.
By managing the disks, file systems,
More recent implementations of
and volumes, it optimizes and simpli-
NAS appliances use SANs to access
fies the delivery of storage to servers
attached block-based storage.
in a network environment, and can
provide services such as redundancy
and replication, relieving servers of
these tasks.

In response to these challenges, in the early years of the 2000s,

storage vendors began to introduce advanced virtualization
features into their products. These features went beyond
utilization improvements, also providing external connectivity
of heterogeneous storage, non-disruptive data migration
and mobility features, enhanced business continuity, logical
partitioning, tiered storage, and thin provisioning. With thin
provisioning (explained further in Chapter 5), the physical
capacity of a volume or file system is allocated when
applications write data, rather than being pre-allocated at the
time of provisioning. Although the need for these capabilities
was well understood by vendors, each vendor had a different
philosophy on where the virtualization should reside.

Drivers for Storage Virtualization

There is no denying that advancements in computer technology
have redefined methods of communication, creativity, and
business management. This revolution has created an explosion
of information generated for both businesses and home users.
In the business world, applications that manage emails and Web

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 10 9/25/09 11:51:45 AM

 Chapter 1: Storage Virtualization: Getting Started 11
sites, as well as customer relationship management (CRM) and
sales systems, are all critical to keeping the doors open. If these
applications are unavailable for whatever reason, businesses
can’t service their customers and they’ll go elsewhere. The
Web has emerged as a source of rich media (images, video,
and audio) and continues to grow unabated, with services
like Facebook, Flickr, and YouTube creating unprecedented
communities and collaboration between people all over the
world. This exponential growth, together with the need to access
information, has placed huge demands on the IT staff looking
after them, especially on their storage infrastructures.

With this explosion of new applications and services, the value

and importance of the information generated has become more
vital. Business intelligence and decision support systems mine
information in databases to create valuable knowledge that helps
organizations make prudent decisions and stay competitive.
Also, with regulatory compliance coming into play, the need to
retain and protect corporate data for predetermined periods
of time is now inscribed in law. Similarly, the accessibility of
information and media from the Internet and Web sites, in
particular its availability, determines how successful a business
is. If users can’t get what they want, when they want it, from
their browser in a couple of seconds, then they simply go
elsewhere.

All this creates several challenges in the storage environment.

Out of storage again!

Perhaps the most nerve-wracking part of a storage
administrator’s job is when he has to go “cap in hand” to the
IT manager or CIO and ask for money to buy more storage. Like
needing to ask your parents for more pocket money, and having
to first explain where your weekly allowance went, system
administrators face a similar judgment. Unfortunately, explaining
“who took all the storage” is not necessarily simple, it just seems
to vanish!

The main reason this happens is applications are typically

allocated more storage than they need. Application and system
administrators often request as much storage as they can get,
rather than what they’re likely to need, since it is a scarce
resource. Twelve months down the track they find they have
used only a small portion of their allocation, say 30 percent.
Due to traditional storage provisioning methods, the unused

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 11 9/25/09 11:51:45 AM

 12 Storage Virtualization For Dummies

70 percent remains “stranded.” Although it hasn’t been used

by the application and is free, it can’t be allocated to other
applications as it is dedicated to the original server. Multiply this
situation out to tens or hundreds of servers and applications
and it’s not hard to see how this inefficiency quickly adds up.

Compounding the problem is the common practice of creating

multiple copies of data. For every byte of production data,
organizations create any number of replicas for backup and
recovery, disaster recovery, data mining, development and
testing. Many organizations exposed to industry or government
regulations retain several more copies for compliance or other
records management requirements long after the data has
become stale.

Countless surveys and assessments of storage capacity

utilization have been conducted over the years, most revealing
that organizations suffer from very poor actual utilization, as low
as 30 to 40 percent.

Reducing cost
As organizations deploy more and more applications into
their environment, with consequent growth of data (often
exponentially), many find they need to buy more storage more
frequently, putting a real strain on budgets. As more storage
is added, without effective storage management tools and
processes, more IT people are needed to keep it going, and
an unending cycle perpetuates. As more allocation requests
are met, more people are needed to monitor the health of the
infrastructure and protect the critical information by ensuring
it is replicated and can be recovered in the event of a failure,
requiring more storage.

With IT budgets and headcount staying flat or in many situations

actually decreasing, organizations need to find more efficient and
automated ways of managing their valuable information assets in
the face of static or shrinking resources.

Managing service levels

IT systems have become the lifeblood of most organizations
in this age. With the dependency on these applications to
service day-to-day business activities, making sure they deliver
consistent and reliable service is critical.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 12 9/25/09 11:51:45 AM

 Chapter 1: Storage Virtualization: Getting Started 13
Take email, for example. Most employees take this service
for granted as a fast way to communicate between their
departments and with customers, send out information or
quotations, and approve business decisions. If the email server
were to slow down or be unavailable for a couple of days,
not only would employees be unhappy, but the organization’s
productivity would plunge to a standstill. In this highly
connected, real-time world, having to resort to traditional forms
of communication and paper-based processes would significantly
reduce the number of transactions the organization could do
and, if the unavailability persisted, could place the business in
real jeopardy. Therefore it is imperative that IT staff monitor
the health of their infrastructure end to end, including storage,
as the key piece to delivering dependable service levels to the
business.

Adding agility and flexibility

With such demanding service levels that require applications
to be available 99.99 percent of the time (or more!), bringing an
application down for servicing and maintenance, or migrating
to new server or storage platforms, would be unthinkable. Try
telling the salespeople in your office that you have to shut down
their quoting and order-entry system for a day and they’re likely
to go after you with pitchforks!

Therefore any such maintenance activities must be done

with the least or, in many cases, no disruption. This poses a
real challenge for IT staff as it typically involves coordinating
multiple people to orchestrate the changes, having sound
processes to execute the tasks with no data loss, and having fall-
back measures in case something goes horribly wrong. All this
sounds like a death-defying high-wire circus act reserved only for
the brave. Many organizations try to avoid such activities as it
places them in a risky situation, but migrations, for example, are
simply unavoidable. Organizations need to look for infrastructure
that is more flexible and agile to cope with these requirements.

In the server world, running out of CPU or memory resources

to meet a given workload is common. With server virtualization
software like VMware, applications and their operating systems
can be seamlessly moved to other physical servers without the
application skipping a beat.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 13 9/25/09 11:51:45 AM

 14 Storage Virtualization For Dummies

For the storage infrastructure, if the storage serving an

application can’t provide adequate performance or availability,
traditionally system and storage administrators would stop the
application, copy all the data to another location, verify it and
hope that the application comes back in good shape. Storage
virtualization technology provides the ability to move data from
one pool of disks to another without any need to do a manual
migration.

Repurposing and adding value

to existing assets
It’s funny how things just don’t seem to last as long as they used
to. Not so much because of wear and tear but because they
become superseded by new models with better performance or
features.

Most storage subsystems and associated networks cost

thousands or even millions of dollars. So, with such potentially
short life cycles, it’s not surprising that organizations are looking
to squeeze every last drop out of their investments.

For many organizations, it’s not a matter of technology being

superseded, but no longer meeting their changing requirements.
Most IT departments assess their storage requirements
according to the most appropriate features, technology,
vendor, and budget to serve their purposes best today.
However, business landscapes are constantly transforming, and
unforeseeable events can change things in a heartbeat. Certain
applications can become more important, requiring greater
protection or performance. Your organization could acquire
another company, and you inherit another IT infrastructure to
manage and merge with your own.

The problem is that business and therefore technology

requirements change rapidly, often within a year, while hardware
may be capitalized in three to five years. If you’ve just bought
a new storage unit and are only one year into a five-year
capitalization cycle or five-year lease, you’re locked out of using
new storage-based technologies for the rest of the capitalization
or lease period.

When organizations compare the cost of maintenance for their

aging hardware with that of buying new hardware, they are
often motivated to buy new. However, with storage virtualization
solutions you can breathe new life into existing assets, rather

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 14 9/25/09 11:51:45 AM

 Chapter 1: Storage Virtualization: Getting Started 15
than replacing them, and repurpose them for a job with fewer
requirements, or extend functionality, such as thin provisioning
or remote replication (these concepts are explained in
Chapter 5), to increase their value to the business.

Going “green”
More than 20 years of research into carbon pollution now
strongly suggests that human activity is having a profound
impact on the Earth’s climate. The almost indisputable evidence
has now forced many governments to action by implementing
tangible strategies and policies that reduce carbon emissions.
As one of the many measures to tackle the problem, it calls for
countries to implement carbon emissions trading schemes to
steadily reduce pollution, but at a cost to businesses.

So what has all this got to do with storage? The net effect
of these new trading systems, coupled with the fact that
the world’s reserves of non-renewable resources are rapidly
depleting, means the cost of power is rising. In the past, the cost
of powering and cooling data centers was negligible and was
often not even factored into the IT budget. However, according
to an IDC article in 2008, recent surveys suggest that with
continued deployment of applications and data doubling every
18 months, power and cooling costs will grow eightfold by 2010!
In another survey, by Gartner Consulting, findings showed that
the power consumption required to run and cool data centers
accounts for almost a quarter of global carbon dioxide emissions
from the information and communication technology sector.
And the StorageIO Group reports that storage itself accounts for
between 37 and 40 percent of total energy usage from hardware.

With business accounting for a substantial share of greenhouse

gas pollution, many organizations have developed social
responsibility policies to help protect the environment.
Demonstrating to customers and their respective industries that
they are not only conscious of the issue, but care about being
part of the solution, is important for organizations. A number
of large multinational companies, such as Citigroup, Vodafone,
HSBC, and British Telecom, have already committed to carbon
reductions within their companies, some aiming to be carbon
neutral altogether.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 15 9/25/09 11:51:46 AM

 16 Storage Virtualization For Dummies

All these factors are prompting businesses to look at innovative

ways to become more efficient users of power in order to
comply with government policies, control their budgets, and
protect their reputations, along with the environment.

Many IT vendors are already actively delivering products

that have less environmental impact. Through streamlining
production processes, many vendors have eliminated waste,
requiring less carbon-dependent power. Smarter designs of
components such as processors and power supplies, and the
ability to spin down inactive disks, reduce the running costs, and
use of non-hazardous materials in the product itself protects the
environment when it is disposed of.

However, these initiatives only go part of the way.

With data storage growing at such exponential rates, together

with the inefficiency of traditional storage networking,
provisioning and application consumption methods, storage
virtualization can go the next step by eliminating unnecessary
waste.

Consolidating storage though intelligent pooling of storage

resources and use of thin provisioning can delay storage
purchases, resulting in far fewer unused disks spinning and
consuming power. By tiering storage resources and aligning the
right storage characteristics to the value of the information,
organizations can use platforms that utilize disk spin-down
technologies for archive data, for example, that is infrequently
accessed.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

03_597705-ch01.indd 16 9/25/09 11:51:46 AM

 Chapter 2

Types of Storage
Virtualization
In This Chapter
✓ Examining storage virtualization throughout the infrastructure
✓ Knowing how to select appropriate storage solutions

I n today’s modern computer architectures, virtualization

exists in almost every layer, from the application to the
operating system, server, networks, and storage devices. For
example, application clustering technologies such as Microsoft
Clusters and Oracle’s Real Application Clusters (RAC) manage
the process of selecting a server to deliver an application
without the user knowing which server it’s coming from. Server
virtualization allows you to run multiple operating systems on
the same physical hardware platform to improve utilization of
the central processing unit (CPU) and memory. Most of these
forms of virtualization work together to optimize efficiency
throughout the layers of technology.

This chapter examines the different forms of storage

virtualization and helps you to choose the appropriate type of
storage virtualization for your environment.

Virtualization can also be implemented in various layers of the

storage infrastructure, starting at the operating systems device
file through to the host bus adapter, storage network, and
storage array, as shown in Figure 2-1.

Virtualization concepts can also be implemented within the

devices themselves. For example, storage array vendors employ
virtualization aspects within devices such as RAID (redundant
array of independent disks). RAID groups together physical

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 17 9/25/09 11:52:12 AM

 18 Storage Virtualization For Dummies

Figure 2-1: Storage architecture follows a logical system.

hard disks, spreading the data across all the spindles and
slicing them up in chunks called LUNs (logical unit numbers)
to deliver to servers. This method provides greater protection
against a physical drive failure, so if one were to stop working
for whatever reason, the system would not lose any valuable
data and continue to work. Similarly, storage network vendors
incorporate partitioning capabilities such as zones and virtual
storage area networks (VSANs) to separate workloads for
management and security purposes.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 18 9/25/09 11:52:12 AM

 Chapter 2: Types of Storage Virtualization 19

Storage Virtualization in Three

Layers of Infrastructure
Modern storage virtualization technologies pool heterogeneous
storage vendor products together in a specific way to provide
advanced features such as non-disruptive migration of data
and thin provisioning (explained in Chapter 5). This level
of abstraction can be implemented in three layers of the
infrastructure, in the server, in the storage network, and in the
storage controller.

In the server
Some of the earliest forms of storage virtualization came not
from the storage infrastructure, but from within the server, or,
more specifically, the server’s operating systems.

With traditional storage hardware devices that connected

directly to servers, the actual magnetic disk was presented to
servers and their operating systems as LUNs, where the disk
was arranged into sectors comprised of a number of fixed-
size blocks. To allow applications to not only store, but find
information easily, the operating system arranged these blocks
into a “file system.” Much like a paper-based filing system, a file
system is simply a logical way of referencing these blocks into
a series of unique files, each with a meaningful name and type
so they can be easily accessed. For example, take the file name
My_Summer_Break.doc. My_Summer_Break describes what is
contained within the file and the extension .doc identifies the
file as a document. This naming is a lot more meaningful than
just a numeric block number.

Although file systems helped to reference information easily, as

more and more of them were created, exhausting the storage
space of the physical LUN, another LUN would be created and
given to the operating system to continue storing files. To know
which data was stored on what LUN, the operating system
would assign each one a volume number, name, or identifier.
In Microsoft Windows you are most likely familiar with a letter
given to each volume such as C:\ or D:\, whereas in UNIX
these look like /dev/hd0 or /dev/hd1. As applications and
users created more files, more volumes were needed to keep up,
pretty soon making it very difficult to manage.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 19 9/25/09 11:52:12 AM

 20 Storage Virtualization For Dummies

Then operating system vendors came up with the concept of

a logical volume manager (LVM). Much like how file systems
grouped blocks together to present files, LVMs grouped volumes
or LUNs together to present larger, more flexible storage pools
to applications, as shown in Figure 2-2. When an LVM started to
run out of space, you could concatenate, or add another volume,
to make it larger without having to reconfigure the application
or shut it down. Conversely, if you had a large volume that
you wanted to slice into smaller chunks, LVMs would let you
partition it in order to separate different information, say one
slice for the operating system itself and the other for user data.

Figure 2-2: Logical volume managers (LVMs) store information in file systems.

Also, since LVMs could hide or abstract where the storage

was actually coming from, you could present multiple physical
disks and perform a “striping” operation. The technique of data
striping takes multiple physical disks, ideally with their own
discrete path to the server, breaks up the files into even pieces,
and spreads the file across multiple storage devices. Writing and
reading data in this way reduces the response time and thereby

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 20 9/25/09 11:52:12 AM

 Chapter 2: Types of Storage Virtualization 21
increases the performance. All hard disks have a seek time,
which is the time it takes for the information being requested on
the spinning disk to arrive under the reading head. By spreading
data across multiple disks, the process of finding the information
can be done in parallel, thereby reducing the seek time.

Although striping greatly helped to retrieve data faster, it did

pose a significant risk that, if just one of the disks failed, then
effectively the full data could not be retrieved, resulting in
corruption. In the 1980s, a new striping method was introduced,
RAID, which increased data reliability if disks failed. RAID-1
duplicates all the data across every disk, providing a backup
copy for everything. Although this provides high levels of
resilience, it is extremely expensive because you need double
the amount of storage. RAID-5, shown in Figure 2-3, combines
the performance of striping plus parity, which provides recovery
in the event of a disk failure. By using a mathematical formula
to calculate the parity, if a disk were to fail in the RAID set,
then the data can be regenerated and retrieved with no impact.
More recently, RAID-6 was introduced to cope with the long
regeneration times of large disk drives, writing parity information
to two disks instead of one.

Figure 2-3: How RAID increases resilience to minimize data loss.

Server-based forms of storage virtualization were originally

incorporated into operating systems as software and still remain
very popular today. Here are the key benefits of this approach:

✓ Server-based storage virtualization is highly configurable

and flexible since it’s implemented in the system software.
✓ Because most operating systems incorporate this
functionality into their system software, it is very cheap.
✓ It does not require additional hardware in the storage
infrastructure, and works with any devices that can be seen
by the operating system.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 21 9/25/09 11:52:13 AM

 22 Storage Virtualization For Dummies

A number of downsides to server-based virtualization also exist:

✓ Although it helps maximize the efficiency and resilience of

storage resources, it’s optimized on a per-server basis only.
✓ The task of mirroring, striping, and calculating parity
requires additional processing, taking valuable CPU and
memory resources away from the application.
✓ Since every operating system implements file systems and
volume management in different ways, organizations with
multiple IT vendors need to maintain different skill sets and
processes, with higher costs.
✓ When it comes to the migration or replication of data
(either locally or remotely) it becomes difficult to keep
track of data protection across the entire environment.

Most operating system vendors like Microsoft, IBM, Hewlett

Packard, and RedHat (Linux) provide at least some capability to
virtualize storage resources. Vendors like Symantec offer more
advanced forms of server-based storage virtualization.

In the storage network

With the introduction of network attached storage (NAS) and
storage area networks (SANs) in the late 1990s, it became
possible to separate disks (and their controllers) from servers
and share the storage resources more effectively among all
applications in the IT environment. The storage network became
the traffic cop for all information being exchanged between
servers and storage devices, and some storage vendors thought
this would be the perfect place to manage virtualization.

Network-based storage virtualization embeds the intelligence

of managing the storage resources in the network layer,
abstracting the view of real storage resources between the
server and the storage array, either in-band or out-of-band.

The in-band approach, sometimes referred to as symmetric,

embeds the virtualization functionality in the I/O (input/output)
path between the server and storage array, shown in Figure 2-4,
and can be implemented in the SAN switches themselves or in
specialized appliances. All I/O requests, along with the data,
pass through the device, with the server interacting with the

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 22 9/25/09 11:52:13 AM

 Chapter 2: Types of Storage Virtualization 23
virtualization device, never directly with the storage device. The
virtualization device analyzes the request, consults its mapping
tables, and, in turn, performs I/O to the storage device. These
devices not only translate storage requests but are also able to
cache data with their on-board memory, provide metrics on data
usage, manage replication services, orchestrate data migration,
and implement thin provisioning.

Figure 2-4: In-band network storage virtualization is embedded in the I/O path.

The out-of-band approach, sometimes referred to as asymmetric,

does not strictly reside in the I/O path like the in-band approach;
rather, it works hand in hand with specific virtualization-
enabled SAN switches to perform specific look-ups, shown in
Figure 2-5. The servers maintain direct interaction with the
storage array through the intelligent switch. The out-of-band
appliance maintains a map (often referred to as meta-data) of all
the storage resources connected in the SAN and instructs the
server where to find it. In this two-step process, the server uses

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 23 9/25/09 11:52:13 AM

 24 Storage Virtualization For Dummies

special software or an agent, as instructions need to be sent

through the SAN to make it work. As data never passes through
the virtualization device, performance is only slightly impacted;
however, functions such as caching of data are not possible.

Both in-band and out-of-band approaches provide storage

virtualization with the ability to:

✓ Pool heterogeneous vendor storage products in a seamless

accessible pool.
✓ Perform replication between non-like devices.
✓ Provide a single management interface.

However, only the in-band approach can cache data for

increased performance.

Figure 2-5: Out-of-band network storage virtualization uses intelligent switches

to maintain direct interaction between the servers and the storage arrays.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 24 9/25/09 11:52:13 AM

 Chapter 2: Types of Storage Virtualization 25
Both approaches also suffer from a number of drawbacks:

✓ Implementation can be very complex because the pooling

of storage requires the storage extents to be remapped
into virtual extents. This requires a mapping table, which
becomes a single point of failure, and a vendor lock-in.
Migration to a different virtualization solution is extremely
difficult or impossible as all the data must be moved back
through the mapping table.
✓ The virtualization devices are typically servers running
system software and requiring as much maintenance as a
regular server. Clustering is needed to protect the mapping
tables and maintain cache consistency between the nodes,
which can be risky. Servers are also limited in the number
of storage ports they can support and RAM (random access
memory) capacity (as storage is limited in cache). As a
result, servers lack scalability.
✓ The I/O can suffer from latency, impacting performance and
scalability due to the multiple steps required to complete
the request, and limited to the amount of memory and CPU
available in the appliance nodes.
✓ Decoupling the virtualization from the storage once it has
been implemented is impossible because all the meta-data
resides in the appliance, thereby making it proprietary.
✓ Solutions on the market only exist for fibre channel (FC)
based SANs. These devices are not suitable for Internet
protocol (IP) based SANs, which utilize iSCSI (Internet small
computer system interface), NAS, or mainframe servers.
✓ Since both approaches are dependent on the SAN, they
require additional switch ports, which involves additional
zoning complexity.
✓ When migrating data between storage systems, the
virtualization appliance must read and write the data
through the SAN, check status coming back, and maintain
a log for any changes during the move that impact
performance.
✓ All of these add complexity and cost to the SAN, which
makes it very difficult to manage the network with
registered state change notifications (RSCNs, which notify
specified nodes of any major fabric changes), inter-switch
chatter, zoning changes, and buffer credit management.
SANs are the third leading cause of application failure after
human and software errors.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 25 9/25/09 11:52:13 AM

 26 Storage Virtualization For Dummies

The in-band approach has a couple of additional disadvantages:

✓ Specialized software needs to be installed on all servers,

making it difficult to maintain.
✓ It also requires specialized SAN switches that support the
propriety virtualization protocol, forcing organizations to
replace existing SAN fabric, which can be expensive.

For all these reasons and more, storage network virtualization

solutions have had limited traction in the marketplace. Vendors
such as IBM, EMC, and FalconStor provide network-based
storage virtualization solutions.

In the storage controller

Enterprise-class storage arrays, which have features and
capability suitable for large organizations, have always featured
virtualization capabilities (some more than others) to enhance
the physical storage resource. One example of this is RAID,
for providing data protection from disk failures. Host storage
domains are another example, for virtualizing front-end
connection ports to allow multiple operating systems to use the
same physical port. Many enterprise-class devices incorporate
sophisticated switching architectures with multiple physical
connections to disk drives to provide balanced performance
and resilience, as well as the ability to logically segment internal
global cache for service quality and security.

With various forms of virtualization already incorporated into

storage controllers for the past 15 years, some storage array
vendors thought it was logical to extend these capabilities to
storage resources outside the box. Controller-based storage
virtualization was born (see Figure 2-6).

This breed of storage array allows other heterogeneous vendor

storage arrays to be directly connected to its controllers. The
external storage assets presented to it are then “discovered” and
managed in the same way as internal disks. This approach has a
number of benefits, including not requiring a remapping of LUNs
or extents, meaning no additional layer of management and far
less complexity in the network. Once virtualized in this manner,
the sophisticated microcode software that resides on the storage
controller presents the external storage assets as if they resided
within the array, with hosts none the wiser of where they are
physically connected.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 26 9/25/09 11:52:14 AM

 Chapter 2: Types of Storage Virtualization 27

Figure 2-6: Controller-based storage virtualization allows external storage to

appear as if it’s internal.

This block-based virtualization effectively enables all

capabilities of the virtualized storage controller to be extended
to the external storage as well. Leveraging mature enterprise-
class features, data can be migrated non-disruptively from one
pool to another, and replication can take place between non-like
and like storage. In addition, partitioning can be implemented
to allocate resources such as ports, cache, and disk pools to
particular workloads to maintain quality of service and security,
and thin provisioning can reduce storage allocation to other
assets. More recently, vendors have introduced clustering
capabilities between controllers to provide the ultimate high
availability in the unlikely event of a complete storage array
failure.

Although a few downsides to controller-based virtualization

exist, the advantages not only far outweigh them but they also
address most of the deficiencies found in server- and network-
based approaches.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 27 9/30/09 10:48:08 AM

 28 Storage Virtualization For Dummies

Here are some of those advantages:

✓ Connectivity to external storage assets is done via industry

standard protocols, with no proprietary lock-in.
✓ Complexity is reduced as it needs no additional hardware
to extend the benefits of virtualization. In many cases the
requirement for SAN hardware is greatly reduced.
✓ Controller-based virtualization is typically cheaper than
other approaches due to the ability to leverage existing
SAN infrastructure, and the opportunity to consolidate
management, replication, and availability tools.
✓ Capabilities such as replication, partitioning, migration, and
thin provisioning are extended to legacy storage arrays.
✓ Heterogeneous data replication between non-like vendors
or different storage classes reduces data protection costs.
✓ Interoperability issues are reduced as the virtualized
controller mimics a server connection to external storage.
✓ All virtualization capabilities can be extended to other
forms of storage and platforms such as NAS, iSCSI, content
archive, and virtual tape library through the connection
of appliance heads. Mainframes can also be connected,
reducing islands of stranded storage. This enables one
common management, data protection, and search
methodology across the entire environment.
✓ Interconnecting legacy assets and non-disruptive data
migration from one platform to another without any
application downtime assists in technology refresh cycles.
✓ Infrastructure can be de-virtualized with full data integrity
since no meta-data is maintained in appliances.
✓ With virtualization embedded in the controller, an
enterprise storage array has ten times the cache size of
a SAN controller and more port connectivity than a SAN-
based virtualization appliance. This enables it to scale well
beyond SAN-based appliances.
✓ A storage controller can enhance the performance of
attached modular storage by as much as 30 percent. SAN-
based virtualization can’t enhance the performance of
external storage and in most cases will degrade it.

Vendors such as Hitachi Data Systems, Sun Microsystems, and

Hewlett Packard employ this form of storage virtualization.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 28 9/25/09 11:52:14 AM

 Chapter 2: Types of Storage Virtualization 29

Choosing the Right Solution

Storage virtualization is an “enabler” in your IT infrastructure;
it helps you maximize the potential of storage assets, creating
efficiencies and automation, and helps avert risky situations. But
how do you figure out which approach is the right solution for
you?

Consider the following issues and how they relate to your

environment to try to identify an appropriate storage
virtualization solution.

What are your objectives?

Understanding the priorities of your business and how the
IT environment can serve to deliver them is critical. If you’ve
suffered from application outage or data loss that had a
significant effect on your organization, exploring how storage
technology can help avoid it happening again is the first place
to start. Or if you find the IT budget allocated by the business is
simply insufficient to meet the growing data requirements, look
for smarter ways of managing your storage resources. Take the
time to match these problems to specific capabilities offered
by storage virtualization. Ask a number of storage vendors to
explain how their solutions can meet your requirements.

Where are you today?

Although the benefits of storage virtualization are enticing and
can make a tangible difference to your business, taking stock
of where your organization is today in its level of maturity and
capability is important. After deciding storage virtualization can
help you in your environment, consider your current capacity
utilization and how much you can save by reallocating storage
resources. Do the vendors you are talking to support the storage
components you already have in your environment? How mature
are your storage management processes and do they need to be
addressed first? Do you have the in-house resources and skills
to undertake such an implementation, or will you need to hire
contractors or rely on your storage vendor? Do you have the
right people to manage the environment after it’s all in place?

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 29 9/25/09 11:52:14 AM

 30 Storage Virtualization For Dummies

Price versus cost

Implementing storage virtualization needs to provide some
measurable benefits to the organization; otherwise, let’s face it,
why would you do it in the first place? Ultimately, stepping into
any storage virtualization solution will require an investment of
time, people, and money. You will need to go through a discovery
phase, assign people to the transition, and possibly buy some
equipment and/or software. However, remember that the price of
all this does not equal the cost. Implemented correctly, the right
storage virtualization system, versus doing nothing, will realize
a return from the investment in efficiency savings. It is therefore
important to project what these savings will be so that you can
evaluate which solution offers the best value. Not all vendors’
solutions will be priced the same and some will offer greater
capabilities that lead to greater savings, so use your organization’s
objectives to prioritize the capabilities that are important. In
Chapter 3, we offer an approach for doing this.

Matching growth to scalability

Not all organizations grow their storage at the same rate, so you
need to evaluate solutions that fit your growth profile. Although
all storage vendors like to talk about the massive scalability and
performance their solutions provide, you need to cut through
the “marchitecture” and only evaluate solutions that best suit
your environment. Understand your real capacity utilization,
your actual application growth (not storage allocation growth),
and shortlist the capabilities you believe will make a difference,
perhaps thin provisioning and non-disruptive data migration.
Insist on keeping to this scope.

Can you afford an outage?

Every storage solution requires some disruption to your
organization while it is implemented. Some solutions require a
greater outage to your environment than others, depending on the
amount of storage you have and the number and quality of skills
you have to carry it out.

As you evaluate solutions, understanding how much disruption

implementing the solution will cause is important. You need to
feel comfortable that the storage or service vendor can minimize
this, not only through the technology itself, but through previous
experience.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

04_597705-ch02.indd 30 9/30/09 10:48:10 AM

 Chapter 3

The Economics of Storage

Virtualization
In This Chapter
✓ Measuring the economic benefits of storage virtualization
✓ Finding the savings in storage
✓ Knowing where to start
✓ Using Hitachi’s Storage Economics model
✓ Looking at storage virtualization in action, a case study

W hen buying a new car, you don’t just consider one

aspect, like the upfront cost of purchasing the car itself;
you take into consideration multiple factors that determine the
overall value for money you are likely to get. You evaluate such
things as how much mileage you’re likely to get from a tank
of gas, what the servicing and insurance costs are, and what
safety features are built in to protect you and your precious
passengers.

The same is true for storage. If the cost of the capacity were
the only factor to consider, then the answer would be simple:
Buy the disk from the guy who gave you the cheapest price per
terabyte!

However, the impact of continuous growth in storage capacity

on the IT budget can be complex, involving the interplay of a
diverse set of costs. If all elements in the storage economics
equation aren’t taken into account, the real cost of a purchasing
decision can be decidedly different than intended.

In this chapter, we first look into these elements and just how
you can measure and evaluate them. Later in the chapter, we
explain Hitachi’s unique Storage Economics model, which

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 31 9/25/09 12:10:30 PM

 32 Storage Virtualization For Dummies

provides a methodology for measuring the benefits, and

introduce a case study to see how one organization realized
significant savings for its business.

How Do You Measure the Economic

Benefits?
Identifying cost categories that are important to the organization
is the first step in assessing the impact of a proposed storage
purchase. We detail these cost categories in the section “Where
Do You Start?” later in this chapter. Once identified, costs must
be quantified using concise, deterministic methods — such as
return on investment (ROI), total cost of ownership (TCO), and
return on assets (ROA), which we examine in this section. Other
methods of evaluation include time values of money saved,
internal rates of return (IRR), and net present value (NPV) of
future savings. These metrics allow differences between new and
old storage infrastructures to be impartially assessed.

Storage virtualization plays an important part in the storage

economics equation as it is an enabling technology to accelerate
efficiency and productivity improvements that justify a business
case for change.

ROI and TCO are two important mechanisms used to determine

the economic value of a proposed purchasing decision. Although
the two are frequently treated as interchangeable in vendor
marketing literature, ROI and TCO have very different purposes.

ROI is effective when challenging the status quo with a

proposition to replace an existing solution. Frequently, IT
planners will have an idea that long-term operating expense
savings can be realized by making an investment, and the ROI
analysis provides the mechanism for fleshing out the financial
pros and cons of the decision. The analysis illustrates how much
is to be invested, how quickly the investment is to be recouped,
and what net savings are to be expected. Figure 3-1 shows a ROI
analysis over three years.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 32 9/25/09 12:10:30 PM

 Chapter 3: The Economics of Storage Virtualization 33

Figure 3-1: This return on investment (ROI) over three years shows a break-
even point a little into the second year.

TCO analysis is used when the financial benefits of two or more

proposed solutions, or a solution against the status quo, must
be assessed. The analysis places the total lifetime operating
and purchasing costs of the assets side by side for comparison
(see Figure 3-2). IT purchasing best practices, documented
by independent analysts, frequently cite the importance of
including a TCO analysis in any competitive bid situation. The
cost categories chosen for comparison are those deemed most
significant, and these frequently vary from one organization to
the next.

Figure 3-2: Total cost of ownership (TCO), comparing the status quo with an
alternative.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 33 9/30/09 10:49:00 AM

 34 Storage Virtualization For Dummies

However, ROA is a broader and more accurate measure of a

company’s ability to make profit from its invested capital. ROA is
the ratio of net income to total assets. Many CxOs often regard
IT storage as having a relatively poor ROA due to the historically
low levels of capacity utilization caused by over-provisioning,
data duplication, RAID (redundant array of independent disks)
overhead, and stranded capacity. With the advent of storage
virtualization, and the ability to repurpose and reclaim legacy
assets, thereby increasing their value, organizations can help
realize a faster return to the business, accelerate the savings,
and reduce the TCO.

So Where Are the Savings in

Storage?
For several years, Hitachi Data Systems has conducted exercises
for hundreds of organizations around the world, assessing the
economic impact of storage on their business. This work has led
to an astounding data point. On average, for every 12 terabytes
(TB) of usable disk capacity within the storage infrastructure,
there is potential for $1-million savings in net operating expense.

This single data point, which is often sufficient to justify further

investigation and economic analysis, is reached through savings
across the business environment. Here are the average savings
in each category:

✓ Reduced waste, 25 percent

✓ Reduced outages, 20 percent
✓ Reduction in labor and management, 15 percent
✓ Reduced maintenance fees, 15 percent
✓ Environmental savings, 10 percent
✓ Miscellaneous operating efficiency improvements, 5 percent
✓ Other savings, 10 percent

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 34 9/25/09 12:10:31 PM

 Chapter 3: The Economics of Storage Virtualization 35

Where Do You Start?

The first step in assessing the impact of a storage system on
the IT budget in your business is to determine what expense
categories are real and defendable to the organization. Operating
expenses (OPEX) fall into two primary categories: hard costs,
the tangible and visible expenses that find their way into
budgets and financial reports; and soft costs, those less tangible
but equally important measures, such as performance and
availability. Each organization will have its own idea of which
costs are hard and which are soft.

Primary areas of investigation should include:

✓ Data protection. Includes the cost of the organization’s

backup and business continuity infrastructure, which
may include tape, replication, mirror sites, and all related
planning and operational activity.
✓ Maintenance. The cost of hardware and software
maintenance is linked to the capacity deployed, whether
in use or not. It may increase as hardware warranties and
bundled service deals expire.
✓ Staff costs. The labor cost of storage administration and
provisioning can be significant, particularly in complex
heterogeneous environments that may have multiple
management interfaces and distributed architectures.
✓ Environment costs. Power, cooling, and floorspace costs
are increasingly important as companies strive to meet
efficiency targets that are often driven as much by a
corporate and social responsibility agenda as the need
to save money. A significant number of companies are
reaching or have reached absolute limits on their data
center growth and power consumption.
✓ Outage costs. These costs will vary widely depending on
the organization type, industry sector, and application
concerned. A realistic estimate of outage costs is a key
element in planning the company’s business continuity
arrangements.
✓ Cost of growth. This will depend on the storage tiering
structure, since each tier is typically subject to different
growth rates, capital expenditure (CAPEX), and hardware
price erosion. For example, lower tiers tend to grow more
quickly than higher tiers and the price of low-cost SATA
(serial advanced technology attachment) hard drives

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 35 9/25/09 12:10:32 PM

 36 Storage Virtualization For Dummies

tends to fall faster than more reliable fibre channel (FC).

Redistributing data between tiers is also a cost.
✓ Information governance. An increasingly important
element in the total cost of storage, with obligations for
companies to hold data for longer periods of time with a
demonstrable search and recovery capability.
✓ Migration. Bringing new capacity online, taking old
arrays out of service, or moving applications between
platforms can involve migration projects that may be time
consuming and expensive, particularly in non-virtualized
environments.
✓ Performance. Extra costs are incurred by the need to
deploy storage systems of sufficient performance and
scalability to meet the requirements of the service level
agreements (SLAs) with the IT users, and thus defer
the need for further investment as user capacity needs
and productivity grow. This may also mean that I/O
performance upgrades are added at the same time as
storage capacity upgrades.

Hitachi’s Storage Economics Model

Hitachi Data Systems, through hundreds of TCO and ROI analysis
studies, has evolved an approach to storage economics and
developed a set of tools and methodologies that can be applied
to a broad range of customer types across different industry
sectors.

The basis of this model compares the existing infrastructure

(business as usual) with a new solution that includes modern
storage innovations such as multiple storage tiers, storage
virtualization, and thin provisioning technologies (explained
further in Chapter 5).

This helps storage managers and CxOs gain a deeper

understanding of the true operational expenses in their
storage infrastructure and establish a roadmap for future cost
reductions. This thorough analysis combines a set of technical
reports with structured customer interviews that provide an
assessment of the storage infrastructure’s operations and costs
in light of the individual business objectives. The methodology
follows three key stages:

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 36 9/30/09 12:44:19 PM

 Chapter 3: The Economics of Storage Virtualization 37
1. Establish a TCO baseline. What does it cost the
organization to own and operate 1TB of storage per
year? The answer is derived from an understanding
of the unique set of costs that apply to that storage
customer, identified in up to 33 different cost areas
ranging from power and cooling to labor charges and
outage costs.
2. Map the costs to a set of strategic and tactical cost-
saving initiatives. What cost-saving initiatives are
needed? These may be strategic or tactical initiatives,
such as consolidation, tiered storage, virtualization,
thin provisioning, chargeback, or implementation of
management tools, and will vary depending on the
customer’s environment. The user maps the costs
identified in the first step to the cost-saving initiatives,
so that the multiple relationships between costs and
investment can be established. This process also helps
to identify the areas of technology investment that are
the highest priorities.
3. Compile a cost reduction roadmap. What will be the
impact on the TCO baseline over the short and medium
terms? With a rollout plan for each of the cost-saving
initiatives, a roadmap can be built that shows the impact
on each of the key cost areas year by year.

By following this three-stage approach, a tactical plan for

addressing the key cost areas and driving long-term TCO
reduction can be scheduled and executed.

Four principles of storage economics

Hitachi has developed an online Go to www.hds.com/economize.
tutorial outlining four principles to
reduce the total cost of ownership
(TCO) with storage virtualization.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 37 9/25/09 12:10:32 PM

 38 Storage Virtualization For Dummies

Case study: Indosat’s storage solution

Indonesian Satellite Corporation Tbk the unique virtualization capabilities
(Indosat), one of Indonesia’s largest of the Universal Storage Platform,
telecommunications providers, was Hitachi Data Systems designed a solu-
formed with the merger of three separ- tion that integrated the three storage
ate entities. The merger transformed infrastructures seamlessly into one
the company into a cellular-focused, environment while leveraging existing
full-network service provider ser- investments. Indosat’s applications
vicing over 12 million customers, and were classified into relevant tiers
presented a major challenge. that matched the value of the data to
the cost of the storage, ensuring the
Working with three different sets of
organization was confident they had a
storage systems threatened to take
sustainable way of optimizing storage
a toll on Indosat’s IT infrastructure
resources that raised efficiency and
operations and maintenance. It was
accountability.
clear that direct attached storage for
each server was no longer an option Virtualization paved the way for con-
as it would not meet Indosat’s require- solidation and allowed Indosat to
ments for scalability and flexibility. make the best use of their existing
storage, effectively protecting the
In addition to this challenge, a rapidly
bottom line. Virtualization also opened
growing customer base was straining
the door to a host of other benefits,
the capacity of the existing systems.
such as reduced management of
A disaster recovery facility was also
provisioning, ease of monitoring to
needed to protect critical application
maintain service levels, and flexibility
data in the event of systems failures.
to conduct data migration and repli-
Consolidation of the IT infrastructures,
cation without increased risk.
as well as the applications, became
essential. The results speak volumes of the
project success: 10 to 20 percent
Indosat identified Hitachi Data
savings in human resources and
Systems as the solutions provider of
40 percent savings in hardware
choice to quickly take control of the
investments. Needless to say, there
situation. By drawing on a combi-
has been no looking back for Indosat.
nation of skills in consolidating
storage resources and leveraging

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

05_597705-ch03.indd 38 9/25/09 12:10:32 PM

 Chapter 4

Making Storage
Virtualization Real
In This Chapter
✓ Understanding key concepts
✓ Planning to virtualize your storage through objectives and analysis
✓ Implementing your virtualized storage solutions
✓ Migrating your data

S torage virtualization can help address many shortcomings

of traditional storage infrastructures, enhance the storage
investment, and respond to real business problems. Walk
into almost any data center with multiple storage platforms
from different vendors and you’re likely to find low storage
utilization, multiple management tools with different interfaces,
and disparate storage management practices for each vendor’s
storage. Storage virtualization solutions help you to simplify
the storage infrastructure and dramatically improve storage
utilization by managing all storage assets as a single pool via a
single management interface.

Data migrations are part of any ongoing IT operation and are

especially intensive during technology refresh cycles. Migration
is highly prone to human error and typically needs to be done
in very small windows of opportunity. A virtualized storage
environment enables an intelligent and dynamic tiered storage
solution that allows seamless data migration across storage
tiers without impacting users or applications. Intelligent tiered
storage solutions help you to simplify infrastructure, ensure
quality of service, reduce risk, and align the right storage tier
to the right application, thereby reducing capital (CAPEX) and
operating (OPEX) expenditure.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 39 9/25/09 11:53:13 AM

 40 Storage Virtualization For Dummies

In this chapter, we deal with the actual implementation of

storage virtualization, but first we outline some of the key
concepts used in storage virtualization.

Key Concepts in Storage

Virtualization
Any technical field develops its own terminology based on
the type of concepts involved, and storage virtualization is no
different.

Here are some of the key ideas to get your head around:

✓ Address space remapping: Virtualization of storage helps

achieve location independence by abstracting the physical
location of the data. The virtualization system presents to
the user a logical space for data storage and handles the
process of mapping it (making it visible) to the physical
location. The actual form of the mapping will depend on
the chosen implementation.
In a block-based storage environment, a single block of
information is addressed using a logical unit number
(LUN); an offset within that LUN is known as a logical block
address (LBA). The address space mapping is between
a logical disk, usually referred to as a virtual disk, and a
logical unit, presented by one or more storage controllers.
The LUN itself may also be a product of virtualization in a
different layer.
✓ Meta-data: The virtualization software or device is
responsible for maintaining a consistent view of all the
mapping information for the virtualized storage. This
information is usually called meta-data and is stored as
a mapping table. The address space may be limited by
the capacity needed to maintain the mapping table. This
capacity is directly influenced by the granularity (the level
of detail) of the mapping information.
✓ I/O redirection: The virtualization software or device
uses the meta-data to redirect I/O (input/output) requests.
The software or device receives an incoming I/O request
containing information about the location of the data in
terms of the logical disk (virtual disk), and translates this
into a new I/O request to the physical disk location.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 40 9/25/09 11:53:14 AM

 Chapter 4: Making Storage Virtualization Real 41
✓ Pooling: The physical storage resources are aggregated
into storage pools, from which the logical storage
is created. More storage systems, which may be
heterogeneous in nature, can be added as and when
needed, and the virtual storage space will scale up by
the same amount. This process is fully transparent to the
applications using the storage infrastructure.
✓ Disk management: The software or device providing
storage virtualization becomes a common disk manager in
the virtualized environment. Logical disks are created by
the virtualization software or device and are mapped to the
required host or server, thus providing a common place or
way for managing all volumes in the environment.

Planning for the Virtualized

Storage Environment
First and foremost, it is important to understand that storage
virtualization is not something you buy off the shelf at your
friendly IT supermarket, plug in, and off you go. Storage
virtualization is a process and, just like any good vacation, the
fun is in the journey as much as the destination. By starting
small, in one part of your environment, you’ll get some benefits.
As you extend storage virtualization to other applications,
steadily the benefits compound, and then you can introduce
more advanced features like migration and thin provisioning to
really accelerate the savings.

After deciding storage virtualization can make a difference in

your environment, planning for its introduction is critical. Much
like when you take a driving vacation, you check that you have
enough oil in the car, check the tire pressure, use maps to know
where you want to go, devise a route, and decide where you’ll
stop for gas.

Every organization is unique, which means that its IT

requirements and priorities are different. Establishing your
objectives and priorities up front is important, because
the success of storage virtualization is largely driven by an
organization’s understanding of its needs. You need to establish
how critical an application is to the business, as well as its
availability and performance. Once the objectives and priorities
are understood, the storage architecture, along with storage
management standards, policies, and operational procedures,

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 41 9/25/09 11:53:14 AM

 42 Storage Virtualization For Dummies

can be devised to meet them. Especially important here is to

identify how storage virtualization can enable you to fulfill those
objectives and priorities.

If the primary objective of the IT organization is to improve the

storage utilization and save on the CAPEX and OPEX by deferring
the storage purchases and decreasing the administration burden,
you need to understand the real scenario in the infrastructure.
What is the current utilization level and how much improvement
does it need? Where are the storage devices and how are they
connected to the servers, applications, and so on? A good
number of SRM (storage resource management) tools work in
either a homogeneous or multi-vendor environment and report
on this. Element managers from respective vendors also give this
information for reference.

If the objective is to reduce the risk and downtime, and improve

the service levels, then you need to do overall performance
analysis of the infrastructure and set the baseline for monitoring
the difference. A good number of tools are available to do this as
well. Unplanned downtime or unexpected outages of the storage
can impact business very badly, so understanding this well is
critical.

Once the objectives and priorities are set, you’re ready to

choose the right method of virtualization (Chapter 2 discusses
different types of storage virtualization to choose from). At this
time, a good strategy is to create a “storage services catalog”
and classify the data storage as required by applications for the
“right” type of storage based on the performance and capacity
needs. We discuss this in more detail in the section “Migrating
Data to a Virtualized Storage Environment” later in this chapter.

Implementing Storage
Virtualization
After you’ve identified the need for consolidation of your storage
environment and planned your storage virtualization, the
next step is to actually make the required logical and physical
changes to the storage environments.

Consolidating and virtualizing storage devices may create

performance issues for the critical applications. Following
best practices to prevent any bottlenecks that may arise and
using the right tools to monitor the environment is important.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 42 9/30/09 10:49:44 AM

 Chapter 4: Making Storage Virtualization Real 43
Partitioning, the subdivision of components such as disks, cache
memory, and ports to configure storage arrays, can assist in
ensuring that these critical resources aren’t overloaded by one
application, which may adversely affect another application.

Partitioning of storage hardware

Virtual partitioning enables logical partitioning of the physical
ports, cache memory, and disks on the virtualized storage
platform. You can then improve security by “restricting access”
so that the administrator of one partition won’t be able to access
or interfere with priorities or data in another partition. Virtual
partitioning also allows you to optimize application performance
within a partition. By dedicating a fixed amount of cache
memory to a partition, you can guarantee that the application
is getting the required resources and won’t be affected by any
other application, host, or user.

Partitioning also enables you to take a departmentalized view

of storage (that is, a view of the common storage pool that
is virtualized). This ability offers the benefits of centralized
management and control, alongside decentralized departmental
control and accountability.

Internal and external storage

configurations
For controller-based storage virtualization platforms, zones need
to be created in the storage area network (SAN) that include the
front-end storage ports of the virtualized storage platform and
the ports of the external storage platforms being virtualized.

Tasks involved in configuring virtualized storage environment

are:

✓ Configuration of external ports. Virtualized storage

platforms can support several types of ports. A port is a
connection point where the storage array is connected
either to another storage device or to the SAN switch.
Target ports (or front-end ports) are used by hosts or SAN
switches to connect to the virtualized storage platform
device. External ports (or back-end ports) are used to
connect to external storage.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 43 9/25/09 11:53:14 AM

 44 Storage Virtualization For Dummies

✓ Discovery of externally attached storage. Using the

defined back-end ports on the virtualized storage platform,
you need to recognize any externally attached storage and
its LUNs.
✓ Discovery of LUNs configured within externally attached
storage. Once storage is discovered, LUNs can be mapped
and managed within the virtualized storage platform’s
software.
✓ Creation of external groups of storage (eGroups) and
external LUNs (eLUNs). This creates LUNs within the
virtualized storage platform, utilizing the defined external
storage.
✓ Mapping of eLUNs to hosts. The new eLUNs can be
allocated to host systems within the environment, just as
for the internal LUNs.

Figure 4-1 illustrates how you can configure a virtualized system

using a virtualized storage platform, with two external storage
arrays.

Figure 4-1: This example of a virtualized storage system uses two external
storage arrays.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 44 9/25/09 11:53:14 AM

 Chapter 4: Making Storage Virtualization Real 45
The sample system shown in Figure 4-1 features the following
characteristics:

✓ Two storage systems (external), ES1 and ES2, are

virtualized behind the virtualized storage platform (VSP).
✓ Three servers, Server A, Server B, and Server C, access
storage from these storage systems.
✓ Storage LUNs from ES1 are migrated onto the VSP’s internal
disks (which are more likely to be high-performance disks).
✓ Server B and Server C access storage via the VSP, either
internally (for high-performance storage needs) or through
storage from ES1 or ES2, which is virtualized.
✓ Server A continues to use those LUNs from ES1 storage
that are not virtualized behind the VSP.
✓ New LUNs will be created within the VSP from its
local disks if Server B and Server C need additional
high-performance storage to support new production
applications. Alternatively, the storage virtualized from ES1
and ES2 can be provisioned to Server B and Server C.

Virtual partitioning can create separate areas and isolate

the effects of the various workloads from one another. For
example, you can define individual storage partitions to isolate
management of one server’s allocated storage from another
server’s allocated storage. You can also create cache partitions
to assign specific cache amounts to the VSP’s internal storage,
ES1, and/or ES2.

Migrating Data to a Virtualized

Storage Environment
After the existing storage is virtualized behind the controller,
you can plan for migration of data from older storage to new
fast-performing storage devices.

Some applications may need very high-performance, highly

redundant storage disks with faster rotation time, whereas
others can use storage from lower performance disks. An
application’s needs may change over time; therefore the ability
to migrate from one type of disk to another on demand is
extremely valuable. Ideally you should perform these migrations
without stopping the applications.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 45 9/25/09 11:53:15 AM

 46 Storage Virtualization For Dummies

Creating tiers of storage is also a good practice, selecting the

right type of disks or LUNs to form a tier for the purpose of
migrating or provisioning. However, locating disks with the right
characteristics from the hundreds and thousands of disks in the
storage environment can be a big challenge.

Classifying the data types

Classification of data is the assignment of different categories of
data to different types of storage media in order to reduce total
storage cost. Categories may be based on levels of protection
needed, performance requirements, frequency of use, and other
considerations. Since assigning data to particular media may be
an ongoing and complex activity, some vendors provide software
for automatically managing the migration process based on a
predefined policy.

Valuable or frequently accessed data, for example, is kept

on high-performance fibre channel (FC) disks, whereas less
valuable data is moved to less expensive near-line storage, such
as a serial attached SCSI (SAS) or serial advanced technology
attachment (SATA) disk, and infrequently accessed data can be
migrated to a high-volume SATA disk or tape. Ideally, utilizing
tiered storage can save money, while easing the access demands
on any single storage tier.

Creating storage tiers

Tiered storage is a strategy deliberately undertaken by an
IT organization to distribute different classes of data across
differently valued storage elements. These tiered systems are
distinguished by differences in cost, performance, reliability, and
feature sets.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 46 9/25/09 11:53:15 AM

 Chapter 4: Making Storage Virtualization Real 47
Tiered storage may also include different classes of disks
within the same storage system, such as 15k RPM (rotations
per minute) versus 10k RPM disks, or FC and SATA disks (on
mid-range systems). A migration of a LUN copies the contents
of the source LUN to a target LUN. At the point of complete
synchronization between the two LUNs, the LUN identifiers are
swapped. Any host accessing the LUN will not see a change
since the LUN identifier did not change.

Although storing everything the same way may appear easier —

on the same class of storage system with the same protection
levels — doing so creates unnecessary costs. The task of
classifying data is somewhat daunting and requires time that
most companies don’t have. A tiered storage approach helps
to move data based on volume groups and their corresponding
application requirements, which relieves companies from the
arduous task of micro-managing the classification of all data.

Uninterrupted data migration

Data migrations are extremely common in storage operations.
Server and storage equipment eventually needs replacement
(typically called technology refresh) due to maintenance
contracts expiring, lease renewals, or no longer being
supported by vendors. Equipment sometimes needs to be
relocated to another data center or consolidated with other
equipment. Other reasons to perform data migrations are to
move infrequently accessed data to lower cost storage systems
or promote data to higher performance storage, as business
dynamics dictate.

Traditional methods of data migration are inherently

disruptive because the process, involving multiple people in
the IT organization (application, system, network, and storage
administrators), can take a great deal of time and resources,
and poses risks if something goes wrong. This is why many
organizations are demanding less intensive ways of managing
data migrations, especially among different heterogeneous
storage tiers.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 47 9/25/09 11:53:15 AM

 48 Storage Virtualization For Dummies

By implementing storage virtualization, heterogeneous migration

of data between one vendor’s storage to another’s is possible.
Defining groups of volumes according to how they relate to the
application and applying policies by way of highly customizable
storage tier definitions enables the movement of data without
interrupting running applications.

Figure 4-2 illustrates the data migrated online from an active

volume from the external storage, ES1, to the tier-one high-
performance internal storage on the virtualized storage platform.
The example shows a man watching a streaming video played on
virtualized storage attached to the server. With dynamic tiered
storage, migration can be executed without any disruption to the
video playing, since it is completely transparent to the server.

Figure 4-2: In this example, the streaming video is uninterrupted during data
migration.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

06_597705-ch04.indd 48 9/25/09 11:53:15 AM

 Chapter 5

Getting the Best from

Storage Virtualization
In This Chapter
✓ Looking at internal storage virtualization through thin provisioning
✓ Understanding management challenges and solutions

R apid storage growth, with its associated requirements

for provisioning and management of storage assets,
has become a major challenge for information technology.
Traditional static provisioning often results in over-provisioning
with significant amounts of unused capacity, which negatively
impacts storage total cost of ownership (TCO). In this chapter,
we look at how internal storage virtualization can improve
both application performance and the utilization of storage
capacity. We also delve into the management challenges faced by
organizations when they implement virtualized storage systems,
as well as the tools and functions required to meet them.

Optimizing Performance and

Improving Storage Utilization
Internal storage virtualization comes in different forms and
capabilities within a storage system. For example, the ability to
abstract all of the disks in large logical pools is a core storage
virtualization technology that provides significant value. Logical
storage pools make it easier to manage and provision your disk
capacity. Additionally, they can provide a huge performance
boost by striping data across dozens, even hundreds, of disk
drives, making all of the disk actuators work in concert to
perform read/write operations. Data striping is a process that
takes multiple physical disks, breaks up the files into even
pieces, and spreads them across multiple storage devices.
These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 49 9/25/09 11:53:40 AM

 50 Storage Virtualization For Dummies

Another valuable internal storage virtualization technology

is thin provisioning, also known as dynamic provisioning. This
technology allows for greater capacity utilization, brings more
applications online without having to buy capacity in advance,
and simplifies the storage provisioning management process. In
this section, we look at various aspects of thin provisioning.

Understanding thin, or dynamic,

provisioning
Thin, or dynamic, provisioning can be used to present a virtual
pool of shared capacity that is larger than the actual amount
of physical storage available. It enables system administrators
to deliver capacity on demand from a common pool of storage.
What differentiates it in the virtualized storage environment
is the ability to dynamically provision from both internal
and external storage, thus allowing advanced virtualization
capabilities on older or lower tier arrays.

The difference between traditional provisioning and dynamic, or

thin, provisioning is shown in Figure 5-1. In the case of a server
that requires 400 gigabytes (GB) today but is projected to grow
to 2 terabytes (TB), traditional methods would require the full
2TB to be allocated today to avoid lengthy downtime for volume
migration or expansion as the data set grows (as shown on the
left). Here the free capacity is not available to re-assign to any
other servers since it is already allocated. Dynamic provisioning
allows a 2TB volume to be presented to the server with only
400GB of physical capacity actually used (as shown on the
right). In this case, 1.6TB of free capacity is actually available to
other servers since it is not allocated. If and when the amount
of capacity approaches 400GB and exceeds defined thresholds,
additional physical capacity can be added online with no impact
to users or applications.

Creating a dynamic
provisioning pool
Dynamic provisioning is implemented using a concept called
a dynamic provisioning pool (DP pool). A large number of
drives can be part of a DP pool, and the storage controller
automatically stripes data across all available disk drives in the
pool. Implementing a pool over a large number of disk drives is
also known as wide striping.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 50 9/25/09 11:53:40 AM

 Chapter 5: Getting the Best from Storage Virtualization 51
Figure 5-1 (on the right-hand side) demonstrates dynamic or
thin provisioning storage pools and thin provisioned volumes.
The host servers see the virtual capacity, represented by the
dotted line disks, whereas the actual capacity is in the DP pool,
represented by the stripes just above the LUN (logical unit
number) parity group.

Figure 5-1: Comparing traditional provisioning with dynamic provisioning.

Benefits of dynamic provisioning

Dynamic provisioning introduces many benefits, such as:

✓ Closely coupling storage purchases to the application’s

consumption.
✓ Repurposing unused storage that is already allocated, using
the storage only when the host (as per the application’s
demand) writes the data on it.
✓ Simplifying storage management tasks by managing
capacity from a central pool instead of server by server.
✓ Reducing capital costs, as over-provisioning is reduced or
eliminated.
✓ Improving performance, as data sets are striped over a
large number of drives.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 51 9/25/09 11:53:40 AM

 52 Storage Virtualization For Dummies

When combined with storage virtualization, the benefits of

dynamic provisioning can be extended to the external virtualized
storage to save space and to improve performance through wide
striping.

Space saving
Local and remote volume copy services that are “thin
provisioning-aware” consume less capacity. With traditional
provisioning, if you have a 2TB volume but only 100GB is actual
data, all of the 2TB will still be copied. In some cases, users
will create two or more local copies of their primary volumes.
With traditional provisioning this would consume lots of
capacity by copying empty blocks; thin or dynamic provisioning
eliminates this.

One technique in dynamic provisioning allows the physical

capacity to be reclaimed if not used by an application, even if
the capacity is allocated to the application. Using dynamic tiered
storage data migration technology, data on a physical “thick”
volume can be migrated to a logical “thin” volume, and unused
allocated space can be reclaimed. This is called zero page
reclaim.

Thin provisioning can enable more applications and servers

per storage system, providing greater levels of consolidation.
Since capacity doesn’t have to be dedicated and fixed on a per-
volume basis with thin provisioning, customers can create more
volumes. As a result, more applications and servers can be
attached to a single storage system.

Due to thin provisioning’s capacity efficiency, it requires fewer

disks, which results in lower power and cooling costs. Often this
is essential — for data centers that are running out of power, for
example.

Wide striping for performance improvements

Mechanically spinning hard drives are the slowest component in
the storage performance chain. The implementation of additional
drives behind each application volume enables more work
to be done in parallel. This is particularly true for massively

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 52 9/25/09 11:53:40 AM

 Chapter 5: Getting the Best from Storage Virtualization 53
interactive applications, including online transaction processing
and email or messaging systems.

This phenomenon is well known by database and storage

administrators, who have been using host-based volume
management software for years to make multiple fibre channel
(FC) LUNs appear to an application as a single volume. Thin- or
dynamic-provisioned volumes defined over a wide-striped pool
of drives can be used to achieve the same performance benefits.

A major benefit of large logical pools of capacity is the ability

to stripe data across a large number of disk drives. This
improves performance significantly, since you can have dozens
or hundreds or even thousands of disk drives conducting I/O
(input/output) operations simultaneously.

Additionally, wide striping significantly reduces performance

management tasks. A great deal of time is often needed
to analyze, tune, and retune performance with traditional
provisioning methods. Wide striping minimizes and potentially
eliminates the need to tune and analyze, since it is perpetually
optimized for performance.

Management Challenges in the

Virtualized Storage Environment
Many environments face ever-growing complexity in the
underlying infrastructure needed to support their core
applications. At the same time, IT organizations are being asked
to do more with lower budgets and the same or decreasing
levels of staff. Therefore, investing in a simplified storage
management infrastructure that supports the goal of doing more
with less is more critical than ever.

Examples of the increased complexity found in today’s

enterprise data center include:

✓ Support for multiple host operating systems

✓ Data storage for multiple terabytes (TB) or petabytes (PB)
of application and file data
✓ Multiple storage systems in the environment, often from
different vendors

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 53 9/30/09 10:50:50 AM

 54 Storage Virtualization For Dummies

✓ Storage, often existing as isolated islands, or “silos,” each

with separate management tools
✓ Application downtime that is unacceptable for routine
tasks (protection, adding capacity, management) on key
production systems
✓ Growing concern regarding data center “hot spots” and
constraints on physical space as capacity increases

Storage consolidation with storage area networks (SANs)

has increased utilization but introduced the likelihood of
performance problems. Applications are more exposed to
performance issues than any kind of component or data center
disaster, so understanding their workloads is very important.
Storage infrastructure is more complex than ever (SAN, cache,
ports, virtualization, thin provisioning). Baselining (analyzing
current performance to provide a baseline for comparison with
other systems) and monitoring can be difficult without the right
tools, but it is critical for implementing storage virtualization
successfully.

Ideally, there should be a minimum set of tools that can manage

and monitor the virtualized storage environment. However, in
reality, IT administrators need to use multiple sets of tools to do
point-in-time management for its various components.

Key points to consider in selecting the right management

software or tools are that each tool should:

✓ Provide end-to-end visibility of the complete storage

infrastructure so the physical connectivity between
applications, server, SAN, the virtualization storage
platform, and external storage is understood.
✓ Collect detailed storage configuration attributes, such as
model, type, status, ports, mappings, RAID (redundant
array of independent disks) groups, and LUNs.
✓ Have the ability to search and filter storage characteristics
(such as size, protection, and performance) to create and
manage storage tier definitions and migration groups.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 54 9/25/09 11:53:41 AM

 Chapter 5: Getting the Best from Storage Virtualization 55
✓ Monitor end-to-end performance of not only application,
server, SAN, and storage, but also virtualization attributes
such as external storage ports, logical partitions, and cache.
✓ Provide detailed monitoring and reporting of thin
provisioning pools (virtual and physical utilization), as
well as the ability to issue alerts when resource thresholds
are met.
✓ Monitor, prioritize, and even cancel migration and
replication tasks to ensure complete control.

Dealing with the challenges

Storage virtualization decreases complexity, allowing one
administrator to manage three to ten times more storage.

With Hitachi Data Systems’ Universal Storage Platform series,

storage virtualization can simplify the management of otherwise
complex storage infrastructures in many ways:

✓ Connectivity and interoperability. Since the Universal

Storage Platform mimics a server host connection to
external storage, interconnecting heterogeneous storage
assets does not require special certification effort or
changes to microcode. Furthermore, since it uses a
standard connectivity protocol (SCSI, or small computer
system interface), it supports more storage systems on
the market.
✓ Unified management. Using a common management tool
set, the Hitachi Storage Command Suite, multiple software
products can be consolidated. In addition, the burden of
management is significantly reduced by using an interface
for provisioning, management, reporting, and protection.
✓ Business agility: Utilizing the power of an intelligently
switched architecture, increased levels of scalability and
performance are available to applications. The platform
also enables non-disruptive data migration to transition
and transform the storage infrastructure to meet changing
business requirements.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 55 9/25/09 11:53:41 AM

 56 Storage Virtualization For Dummies

Managing performance and

capacity
Earlier in this chapter, we discuss the new challenges
IT organizations potentially face when introducing new
technologies such as virtualization and IT consolidation, as well
as the management tools and functions needed to deal with
those challenges. Performance and capacity management of the
virtualized environment is also essential.

With the ability to partition storage resources into multiple

virtual machines, each with its own dedicated external capacity,
cache, and ports, you can minimize the chances of a rogue
application consuming all the resources and impacting other
applications.

Furthermore, since each virtual storage machine looks like a

separate, isolated storage device, IT teams can associate specific
business applications with selected departments and business
units, then charge back accordingly. These capabilities allow
organizations to create a utility model to deliver metered storage
resources to select business entities. Subsequent management of
virtual storage machines can be addressed centrally or by local
administrators.

Measuring key metrics when applications and infrastructure are

running well is an important task for maintaining service level
agreements (SLAs). These baselines serve as a reference point
to continually assess the infrastructure against. Any resources
(CPU, bandwidth, storage) that have the potential to impact
application performance, and therefore the service level, can
be easily identified. Once recognized, thresholds can be set and
alerts defined to ensure that the IT organization can act upon
them before a problem arises.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

07_597705-ch05.indd 56 9/25/09 11:53:41 AM

 Chapter 6

Ten Best Practices

for Deploying Storage
Virtualization
In This Chapter
✓ Putting in place a process for establishing physical configurations
✓ Choosing the right way to implement data migrations and tiered storage
✓ Getting the most from thin or dynamic provisioning
✓ Discovering scenarios in which storage virtualization can improve IT
efficiencies

I n this book, we give you a pretty good primer on storage

virtualization — the concepts, benefits, and different types.
We examine the value of storage economics to validate the
savings that storage virtualization can bring. We also look at how
to plan for different techniques and methods used in storage
virtualization, such as data migration, tiered storage, and thin, or
dynamic, provisioning.

In this chapter, we discuss a design methodology and best

practices for developing and deploying the storage virtualization
technologies.

#1: Choose the Right Storage for

the Host’s Mapping
In order to implement storage virtualization for physical
configurations, as a first step, you need to gather information

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

08_597705-ch06.indd 57 9/25/09 11:54:00 AM

 58 Storage Virtualization For Dummies

on all host systems and disk configurations. Then allocate the

disks to the hosts from the storage subsystems by discovering
them from the host bus adapter (HBA). To ensure that the target
logical unit numbers (LUNs) are available on the storage area
network (SAN), the best practice is to use an HBA tool (such as
HiSat from Hitachi Data Systems) to locate them. This way you
know that the mapping is done properly. In certain cases, this
process may require a reboot.

Having multiple paths for storage resources is always

recommended for load balancing and failover (automatically
switching to a different redundant system in the event of
failure) or redundancy. Increasing the number of paths from the
virtualized storage platform to the external storage results in a
significant increase in the speed of data transfer. Updating I/O
(input/output) activity will have a major impact.

#2: Assign External Storage

Configurations
After selecting the right storage for the host’s mapping (refer to
the previous section), the next step for physical configurations
is to assign external storage groups (eGroups). You also need
to assign external LUN (eLUN) addresses to the LUNs that have
been discovered by the virtualization layer on the virtualized
storage platform for the external storage systems.

Disabling cache (cache mode) while assigning eLUNs for data

movement purposes is very important. Having cache enabled
forces the use of the virtualized storage platform cache for
updates. If, at the time of de-virtualization, all segments from
cache were not de-staged, data corruption occurs — which can
result in serious customer data integrity issues. When creating
host storage mapping, be sure to specify the right host modes
for the server that will be connected (such as Windows or
UNIX).

#3: Format the Virtual Disk

The next step in implementing storage virtualization for physical
configurations is to use the virtualized LUNs instead of the
native LUNs. To do this, you need to delete the partitions from
the native LUN and re-create the partitions using the operating

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

08_597705-ch06.indd 58 9/25/09 11:54:00 AM

 Chapter 6: Ten Best Practices for Storage Virtualization 59
system command (such as the Solaris format). When virtualizing
an existing LUN, a good practice is to run the format command
in order to display the partition table of the native volume.
Doing this will allow you to re-create the partitions on the new
virtual LUN using the right geometry.

After virtualizing the storage, the host may see two disks for
the same physical device, one for the virtualized device (target)
and the other for the original device (source). Understanding
the source device in the controller–target–device (such as
c1–t1–d1) combination and correlating it with the target device
(say, c2–t2–d2) is critical after the virtualization and volume
migration. This will help keep track of the actual storage model
and device, and keep the integrity of the data that resides on
the disk.

Do not use the Auto Configure option during the formatting as

this can modify your disk label.

#4: Choose the Right Disk

Attributes for Data Migration
After the physical connections and host storage mappings are
defined (refer to the previous sections), you’re ready for data
migration. This step is about migration of data from a volume
in one storage subsystem to another. For the migration to take
place, a target volume with the right compatibility must exist.
Key criteria are size and emulation mode.

When performing the migration of volumes using a virtualization

storage platform that has other workloads running, the best
practice is to use a dedicated virtual storage partition. This way
the migration will have minor impact on the other workloads.

#5: Migrate the Data

Use the following guidelines to achieve an effective migration
process:

✓ Do not perform migrations when the virtual storage

platform’s cache write pending rate is higher than
60 percent.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

08_597705-ch06.indd 59 9/25/09 11:54:00 AM

 60 Storage Virtualization For Dummies

✓ Make sure that the updating I/O rate is below 50 IOPS for
the source volume.
✓ Do not perform multiple migrations targeting the same
RAID (redundant array of independent disks) or array
group as this could adversely affect performance.
✓ Monitor the RAID group busy rates while selecting the
target LUNs for migrations.
✓ Do not perform migrations during the following operations:
• Backups or restores
• Application upgrades
• Month-end processing
• Quarter-end processing
• Year-end processing
• Payroll cycle

#6: Classify the Data

The design of a virtualized tiered storage system should start
from the application’s perspective. The business needs and
applications drive the storage requirements, which in turn
guide tier definition. Most applications can benefit from a mix
of storage service levels, using high-performance and less
expensive storage depending on the usage and business needs.

Recommended practice is to develop a storage services catalog

of tiers with predefined characteristics and then allocate
storage to applications, as needed. Figure 6-1 shows a multi-tier
model outlining main classes of data tiers — production data,
protection data, and archival data. Production data supports
your primary applications, whereas protection data stores
copies of data for recovery. Archival data is for stale but still
important data (possibly on lower cost storage) for compliance
reasons. Your individual requirements may call for more or less.

Most customer environments contain a mix of workload and

applications within a single system, so when you design the
storage system make sure that I/O load or activities (such as
formatting, shredding, replicating, or parity rebuilding) on a tier
do not disrupt or degrade other tiers. In particular, don’t allow a
lower tier load to affect higher tier performance.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

08_597705-ch06.indd 60 9/30/09 10:51:57 AM

 Chapter 6: Ten Best Practices for Storage Virtualization 61

Figure 6-1: Types and attributes of data and storage tiers.

#7: Understand the Workload or

I/O Profile
In looking at performance requirements for designing or
selecting a storage tier, understanding the application’s use
of that tier’s storage is also important. The mix of read/write
characteristics is a major determining factor in the storage
performance it will see.

Ask yourself:

✓ How much is a sequential read?

✓ How much is a random read?
✓ How much is a sequential write?
✓ How much is a random write?

Consider an I/O’s data path inside the virtual storage platform

when answering the preceding questions:

✓ Random read hits are serviced at fastest speed by cache

and do not reach the back end.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

08_597705-ch06.indd 61 9/25/09 11:54:01 AM

 62 Storage Virtualization For Dummies

✓ Random read misses go through cache unaltered and go

straight to the appropriate back-end disk drive. This is the
only type of I/O operation where the host always sees the
performance of the back-end disk drive.
✓ The host sees random writes completed at the fastest
speed; it only sees delay if too many pending writes build
up. Each host random write is transformed when going
through cache into a multiple I/O pattern that depends on
the RAID type.
✓ Host sequential I/O reads and writes are done at fastest
speed. The cache acts like a holding tank and the storage
system back end puts (or removes) “back-end buckets”
of data into (or out of) the tank to keep the tank at an
appropriate level.

High-performance drive type and RAID configurations will have

a much bigger impact on a random read-and-write–intensive
application, versus one that does mostly sequential reads and
writes. Proper cache sizing, I/O queue depth settings, and
cache partitioning can be equally important as drive type and
configuration in tiered storage designs.

Always use the right software tools that can interface to and
are supported by the platform’s firmware, and that provide
enhanced levels of functionality to optimize and tune all drive
and RAID combinations.

#8: Design the Dynamic

Provisioning Storage Pool
To gain optimum performance and benefits using thin
provisioning:

✓ Design the thin provisioning configuration with multiple I/O

and/or purpose-specific pools.
✓ Understand that multiple pools may conflict with the
goal of widest possible striping and simplified storage
provisioning (via less pools to manage).
✓ Validate application requirements carefully through
baselining. Think of each pool as a tier and use data
classification for proper assignment.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

08_597705-ch06.indd 62 9/25/09 11:54:01 AM

 Chapter 6: Ten Best Practices for Storage Virtualization 63

#9: Create a Dynamic

Provisioning Pool
When creating the thin or dynamic provisioning storage pool:

✓ Do not intermix the hard drives with different

characteristics.
✓ Create the pool with a minimum of four RAID groups, with
one LUN in each consuming all of the RAID group’s usable
space. Each RAID group should be behind a different back-
end director pair.
✓ For dynamically adding more storage to the thin
provisioning storage pool, use normal I/O capacity planning
to determine the size of expansion.

#10: Plan and Design a

Virtualization Project
Successful implementation of a virtualized storage environment
and advanced techniques (such as a tiered storage strategy and
thin provisioning) will provide significant business advantages
and an improved return on investment (ROI).

Always use the professional services offered by respective

vendors, because they know their technology, hardware and
software well, and they can design, plan, and implement the
project successfully. They should also teach you how to manage
the project after the implementation is completed.

Through proper implementation of storage virtualization,

IT can be more accountable and will provide business-level
reporting on a common set of metrics. Many companies do
not measure performance in a consistent way, especially with
newer virtualization technologies. Few have good cross-domain
coordination or have established an agreed-upon set of metrics
for monitoring and managing performance expectations. Service
level agreements (SLAs) can’t be implemented without a good
understanding of the metrics.

If the right metrics are put in place, with tools and processes to
help measure and manage them, then performance SLAs can be
established relatively easily.

These materials are the copyright of Wiley Publishing, Inc. and any
dissemination, distribution, or unauthorized use is strictly prohibited.

08_597705-ch06.indd 63 9/25/09 11:54:01 AM