Gartner Reprint(Original)
Uploaded by
blackzigoraitGartner Reprint(Original)
Uploaded by
blackzigoraitLicensed for Distribution
How to Evolve Your Physical Data Center
to a Modern Operating Model
Published 29 March 2022 - ID G00749388 - 11 min read
By Jason Donham, Jonathan Forest
As enterprise infrastructure moves toward cloud distributed
application delivery, the data center itself is being transformed. I&O
leaders must transition to a hybrid model mindset where the
presence of an on-premises data center is no longer the primary
driver for infrastructure decisions.
Overview
.
Key Findings
No longer a physical location, the data center infrastructure is evolving
into the new IT operating model brought about by hybrid cloud.
Enterprise operational and deployment models are reaching beyond
public cloud to include hybrid cloud-based platforms for on-premises
as-a-service offerings.
Acquiring and sustaining skills is proving challenging as the speed and
scale of modern application deployments require infrastructure
automation to allow I&O to keep up.
Recommendations
I&O leaders deploying data center infrastructure should:
Expand deployment options by integrating hybrid cloud into a holistic
architectural approach for application delivery.
Implement core data center infrastructure functions by leveraging as-a-
service models to deliver performance, security and agility at the right
price point.
Enhance their ability to meet customer demand by investing in
infrastructure automation, skills and tools.
Strategic Planning Assumptions
By 2025, 40% of newly procured premises-based compute and storage
will be consumed as a service, up from less than 10% in 2021.
By 2025, 70% of organizations will implement structured infrastructure
automation to deliver flexibility and efficiency, up from 20% in 2021.
Introduction
Placement of application workloads is driving infrastructure decisions
rather than infrastructure determining application location. Hybrid
cloud, defined as public cloud, edge, colocation and on-premises
locations, is rapidly emerging as an alternative to public cloud for
mission-critical workloads and as-a-service offerings. Data center
functions are no longer centralized in a physical location, but rather
deployed to meet complex business requirements by utilizing cloud,
data center, colocation and edge deployment locations (see Figure 1).
Figure 1: Your Data Center Is No Longer Limited by Walls
I&O leaders are finding the data center more difficult to design and
manage as workloads and infrastructure expand beyond traditional
centralized locations. To solve these complex issues, a wider view of
data center architecture and operations must be taken. I&O leaders
must also automate the data center environment and acquire
automation skill sets. Proper leveraging of as-a-service models will also
ease the infrastructure management burden by outsourcing certain
responsibilities to partners.
This research provides I&O leaders with a path forward to address the
separation of data center and data center infrastructure.
Analysis
Expand Deployment Options by Integrating Hybrid Cloud Into
a Holistic Architectural Approach for Application Delivery
I&O leaders must understand the effects of data center infrastructure
distribution across their hybrid cloud. The physical data center footprint
continues to be more concentrated and shrink as applications and
services migrate to cloud-based platforms. High availability of
applications, the desire for conversion from capital expenditure (capex)
to operating expenditure (opex) and fewer available technical resources
are driving applications out of the traditional data center.
The functions of the data center no longer reside within the traditional
four physical walls of a single building per se. However, the entirety of
the infrastructure must be viewed as one hybrid cloud IT operating
model within virtual boundaries for planning and administration
purposes. Infrastructure can no longer be viewed as individual
components at the domain level but must be viewed strategically at the
system level. This holistic approach ensures proper end-to-end design
and management that will lead to efficient, simple, agile and secure
operations.
Data center functions, including networking, backup, compute and
storage, are becoming increasingly virtualized. I&O leaders must
leverage this capability and place data center functions in support of
workloads where they will be most effective to their business in terms
of cost, performance, agility and security (see Tool: Workload
Placement in Hybrid IT). In many cases, this will require multiple
iterations of the same function in different locations to meet the
requirements of distributed applications.
In a distributed application deployment, the end-user experience is
typically the driver for the location of application components. However,
data sovereignty brings its own unique set of challenges as data now
resides in more locations throughout the enterprise. The same dataset
may be subject to different laws depending on where it is collected and
located. The repercussions of data crossing international borders must
be accounted for in a distributed computing model where data is often
traveling from one part of the enterprise to another.
I&O leaders must account for regulatory requirements, and ensure that
the data location and management control plane are integrated into the
decision-making process. Only data that is required to be processed
locally should be placed in locations that have restrictive regulations to
ensure that organizations can use their data effectively.
Flexibility, simplicity and ease of management are key attributes for
system design in a distributed environment. Taking a macro view of the
enterprise, the overall systems architecture must be designed with
flexibility and scalability in mind so that it can support the changing
needs of end users and applications.
Ease of management is accomplished through standardization where
possible, along with robust toolsets that will allow monitoring and
control of disparate systems. On-premises systems must contribute the
same flexibility, simplicity and ease of management to remain a viable
location for distributed systems. There is a delicate balance that must
be maintained because the distribution of functions adds increased
complexity, security risks and fragility due to more moving parts and
interdependencies. However, distribution can add agility and resilience
if the system is properly designed and implemented.
Most organizations exist somewhere within a hybrid cloud model, in
between the legacy on-premises data center and public cloud extremes.
This blended environment of traditional, public and private cloud
services will be the primary operating model out of necessity for at least
the near term for most organizations because not all applications are
able to be migrated to the public cloud. In fact, Gartner estimates that
only 50% of workloads will move to the cloud by 2025 and that only one-
third of applications will remain on-premises beyond that.
Organizations choose to retain some applications on-premises for a
variety of reasons including:
Security concerns
Custom-built application with architecture that is not cloud-compatible
Data sovereignty and regulatory demands
Legal concerns
Data residency for latency concerns
Protecting intellectual property (IP) from transfer
I&O leaders must embrace the hybrid model by leveraging the
efficiencies that can be gained through intelligent infrastructure-aided
placement of the workloads they support. Hybrid model architecture is
agile and flexible in order to deliver a wide range of deployment options
to support the varying needs of applications (see The Disappearing
Data Center Opportunity for Infrastructure and Operations Leaders).
Implement Core Data Center Functions Using as-a-Service
Models
As-a-service models are common in public cloud deployments due to
ease of scale, pay-for-consumption pricing and the elimination of
hardware asset ownership. There is a growing interest in and adoption
of these models on-premises as well. On-premises consumption-based
models place hardware and software on-site and utilize different levels
of third-party management, depending on which service type is utilized.
Services may come from multiple sources and may now be both cloud-
based and on-premises. This results in more “cloudlike” benefits on-
premises. Applications may be designed and deployed on-premises
using cloud development tools, resulting in greater portability, better
security, more cost-savings and faster time to market.
On-premises consumption-based infrastructure provides flexibility for
workloads not able to move to the public cloud. These platforms are
being driven by hardware providers as the platform of choice for private
cloud and hybrid cloud models. Benefits include potential cost savings
from capex to opex conversion, resource scale, burst capability without
purchasing additional infrastructure and infrastructure automation (see
Market Guide for Consumption-Based Pricing for Data Center
Infrastructure and Enterprise Storage as a Service Is Transforming IT
Operating Models). The consumption model is a good fit for businesses
requiring flexible resources along with outsourcing of hardware
ownership and infrastructure management (see Figure 2).
Figure 2: How Pay-per-Use Infrastructure Consumption Models
Work
As-a-service models have broken the traditional purchase/depreciation
life cycle. Organizations can now procure only what they need for the
period of time they need it and avoid the capital expenses of owning the
infrastructure and disposing of it at the end of its useful life. The cloud
as-a-service model is more flexible due to being driven by software in a
multitenant environment. On-premises services are somewhat limited
by contract length and a single tenant hardware deployment even when
considering the availability of burst capacity.
The ability to quickly scale resources up or down to meet usage
demand for cost-effectiveness is important due to the challenges of
positive or negative growth and capacity planning inherent in the
distributed model. The capex model is very rigid and therefore
unsuitable for the modern “on-demand” needs of developers and end
users.
Many organizations do not have the in-house skill sets or resources to
manage their infrastructure through a refresh cycle and redesign
process. Others simply do not have the desire to manage their
infrastructure as it is not a core component of their business model.
With as-a-service models, resource allocation or capacity expansion
becomes a service request with a predetermined service-level
agreement (SLA) rather than a drawn-out process of budgeting, design,
procurement and deployment. As-a-service models allow businesses to
focus on outcomes without long-term capital commitments.
Risk mitigation strategies can also benefit from the as-a-service model.
For example, as a computing environment becomes more distributed
and technically complex, it becomes progressively more difficult to back
up files and virtual machines. The backup scheme becomes more
difficult to manage as the backup systems themselves become more
complex. Hardware and software obsolescence are no longer an issue
as the service provider assumes all responsibility for backups, testing,
restoration of data and hardware refreshes.
Increase Your Ability to Meet Customer Demand by Investing
in Infrastructure Automation Skills and Tools
Infrastructure automation may be accomplished through platform as a
service (PaaS) offerings that are deployed with vendor-supplied
automation software. Automation can also be achieved through off-the-
shelf software designed specifically for enterprise environments. In
either case, routine tasks can be offloaded to the software automation
suite in order to provide operational efficiency and improved
productivity by freeing your IT teams from tasks that can be completed
without human intervention. In a recent Gartner survey, I&O leaders
indicated the need for enterprise automation in their data centers (see
Figure 3).
Figure 3: Over Half of I&O Leaders State Their Data Center
Infrastructure Most Urgently Requires Automation
Artificial intelligence operations (AIOps) are being integrated into
automation solutions to improve agility, ease of scale, error reduction
and overall simplification of processes. Automated analytics monitor
the infrastructure and act to manage and maintain it much faster than a
human operator is capable of doing. For example, storage vendors in
particular have fused AIOps into their products to the point that much of
the traditional storage administration and support overhead has
disappeared. AIOps self-management capabilities drive down costs and
enhance agility. We expect this technology to become more prevalent in
enterprise networks over time to minimize outages and accelerate
trouble ticket resolution time. Faced with increasing distribution and
complexity of systems, I&O leaders are finding their IT teams lack the
skills necessary to manage such an environment.
To improve the scope and scale of infrastructure automation, I&O
leaders must:
Acquire/develop/sustain automation skill sets across multiple
technology domains.
Encourage an “automation first” mindset across their teams.
Drive collaboration between technology teams to remove handoffs and
wait states that slow down delivery.
Training and skill set acquisition must be focused on this team to
ensure the right tools and the right personnel are in place to manage the
enterprise effectively. Automation technology skills such as Python
scripting, Ansible, Puppet, Chef and Terraform are examples of useful
skills for teams to master.
It is important to focus on matching the tools deployed to the maturity
of the organization, focusing on task automation with a standard set of
tools, and evolving to a cross-domain process automation capability.
This keeps management simple and efficient enough for a single team
to accomplish. As more infrastructure becomes automated, internal IT
teams must also be able to effectively communicate with vendor-
partner teams that will be managing some aspects of the environment
through services.
Evidence
Gartner analysts conducted more than 6,000 client inquiries on the
topic of data center infrastructure during 2021.
The 2021 Gartner I&O Leaders Survey was conducted online from 14
through 25 June 2021 to track burning issues for infrastructure and
operations (I&O) leaders and where they were prioritizing their
investments over the next year. It also sought to explore investments in
cost optimization and innovation strategies. In total, 96 Research Circle
members participated, of which 71 were IT leaders, 22 were CIOs, two
were CSS leaders and one was a CFO. Survey respondents were from
North America (43), EMEA (33), Asia/Pacific (10) and Latin America (9)
responded to the survey.
© 2023 Gartner, Inc. and/or its affiliates. All rights reserved. Gartner is a registered
trademark of Gartner, Inc. and its affiliates. This publication may not be
reproduced or distributed in any form without Gartner's prior written permission. It
consists of the opinions of Gartner's research organization, which should not be
construed as statements of fact. While the information contained in this
publication has been obtained from sources believed to be reliable, Gartner
disclaims all warranties as to the accuracy, completeness or adequacy of such
information. Although Gartner research may address legal and financial issues,
Gartner does not provide legal or investment advice and its research should not
be construed or used as such. Your access and use of this publication are
governed by Gartner’s Usage Policy. Gartner prides itself on its reputation for
independence and objectivity. Its research is produced independently by its
research organization without input or influence from any third party. For further
information, see "Guiding Principles on Independence and Objectivity."
About Careers Newsroom Policies Site Index IT Glossary Gartner Blog Network Contact Send
Feedback
