You are on page 1of 11

Network Managers' 10 Most-Challenging

I&O Trends
19 October 2010
Jay E. Pultz
Gartner RAS Core Research Note G00206017

Gartner has identified 10 key infrastructure and operations trends that will significantly affect network design
and planning during the next three years. Recommendations are made to deal with each of them.

Overview

Enterprise network managers must factor into their plans and designs 10 key infrastructure and operations
(I&O) trends that will significantly impact networking during the next five years. This research explains these
trends, recommends specific actions and provides a timeline for network managers to use as a guide.

Key Findings

The 10 I&O trends including significantly increased bandwidth, improved service quality,
rearchitecting network designs, reducing latency and more closely aligning the network operations
center (NOC) with IT operations will have a heavy impact on networking.

Network managers can no longer design and manage networking in a vacuum, because these I&O
trends will require considerable interaction within I&O, with the rest of IT and with the business itself.

New network cost reduction opportunities are more likely to be derived from design optimization than
from negotiations for rate reductions or utilization of telecom expense management services.

Recommendations

Because these trends significantly affect the network, network managers should be involved in all
phases of the rollout of key I&O initiatives in their enterprise, including planning, implementation and
operations.

To help fund initiatives, focus on lowering telecommunication service provider (TSP) costs in the
short term, but with network optimization in the midterm to long term.

Reorganize the NOC around processes versus technology platforms; consider integrating NOC
processes with those of IT operations.

Test or emulate the performance of cloud-based applications in all geographies where you plan to
deploy them. Directly interconnect networks used in cloud-based services, and use application
delivery controllers (ADCs)/WAN optimization controllers (WOCs), where necessary.

Table of Contents

Analysis

Cost Optimization
Consolidation
Virtualization
Convergence
New Client Architectures
ITIL
Operations Integration
IT (and Network) Services
Automation
Cloud Computing

List of Figures

Figure 1. Ten Key I&O Trends for Network Managers


Figure 2. I&O Networking Costs
Figure 3. Network Cost Reduction Actions
Figure 4. The ITIL Framework for ITOM Processes
Figure 5. NOC Transition
Figure 6. Options, Characteristics and Examples of Use for Network Cloud Computing

Analysis

This research focuses on 10 key trends in I&O outside of networking about which network managers need to be
aware. They should be ready to alter their plans to accommodate them. These trends will have a high impact on
networking, with some requiring significant additional bandwidth and a rearchitecting of planned network designs.
Although the concepts behind most of these trends are not new, what can be implemented during the next several
years will affect networking more significantly than ever before. Network managers can no longer design and
manage networking in a vacuum, because these trends will require considerable interaction with other I&O
leaders, including data center managers, the rest of IT and the business itself. Network managers should meet
with their I&O peers to create a plan of action that will involve interaction within and among groups.
To a large degree, how you address these trends determines your future. The time to act is now. This research
discusses each trend, how it affects networking and what actions network managers should take to address them.
Figure 1 lists the 10 trends; two relate to I&O, and four relate to either I or O (that is, either infrastructure or
operations). Figure 1 also arrays these 10 trends in time in most enterprises, aspects of these trends are either
being implemented or will be under way shortly. The time sequencing here suggests an overall prioritization;
however, network managers need to also factor in specific I&O plans in their enterprises to react to these trends.
Figure 1. Ten Key I&O Trends for Network Managers

Source: Gartner (October 2010)


Back to List of Figures
Back to Table of Contents

Cost Optimization
I&O comprises about 60% of annual enterprise IT spending (see Figure 2). Gartner's annual CIO survey indicates
that cost reduction is the CIO's second highest priority, behind better aligning IT with the business. Hence, I&O
leaders including network managers are under pressure to reduce their costs. As shown in Figure 2, the data
center and networking claim the lion's share of I&O costs; networking is conservatively more than 20% of an
enterprise's total IT expenditures, excluding wireless. For most enterprises, carrier and staffing costs dominate
networking budgets, together comprising about 80% of the total. Hardware, software, maintenance services,
occupancy and a small percentage of unallocated costs make up the remainder.
Figure 2. I&O Networking Costs

Source: Gartner (October 2010)


Back to List of Figures
Back to Table of Contents

Because nearly half of the network expenses go to TSPs (such as Verizon and BT), network managers need to
continue to renegotiate contracts with these vendors to ensure their contracted rates are market-based. In the
past, attention tended to be focused on voice calls (cost per minute) and data WAN costs, such as cost per
Multiprotocol Label Switching (MPLS) port. Access remains a major cost; clients should be looking to drive
metropolitan Ethernet costs down. Additionally, compared with 2009 rates, we expect managed service costs to
plummet by 50% or more by 2012.
However, longer term, network cost savings will derive heavily from optimizing network design. Gartner's
recommendations here include:

Using a virtual private network (VPN) rather than MPLS for smaller sites
"Right sizing" network availability
Simplifying network designs e.g., reducing the number of tiers in LANs
Competitively sourcing network equipment, such as Ethernet switches and IP routers
Using Session Initiation Protocol (SIP), instead of Integrated Services Digital Network (ISDN) Primary
Rate Interface (PRI) trunking

We cannot go into detail here, but we want to point out primarily that:

Although contract renegotiations, especially with TSPs, will still be a major contributor, clients should
further investigate opportunities to refine the design and sourcing of their networks.

Substantial steps can still be taken to optimize network costs (see Figure 3).

Figure 3. Network Cost Reduction Actions

Source: Gartner (October 2010)


Back to List of Figures
Back to Table of Contents

Consolidation
In the past, the rise of distributed computing helped to drive the decline of the large data-processing sites that
characterized the era of the mainframe. Today, data centers are rising in importance. Server rationalization,
hardware growth and cost containment trends drive the consolidation of enterprise data-processing sites into
larger data centers.
Changes in networking have made it more viable to consolidate data centers: bandwidth costs have declined
dramatically (10% annually for more than a decade), broadband is now widely available over vast areas and truly
diverse access options are now available Nonetheless, networking represents a large cost to make data center
consolidation happen.
With servers located in the data centers, instead of at remote sites, higher bandwidth is now required for the data
center. This bandwidth needs to be highly available, because business at the site may be affected if the servers
cannot be accessed. Lastly, higher class of service options for WANs, ADCs and WOCs may be required to achieve
the desired response times. Network managers should be involved early on in these projects to fully ascertain the
networking requirements here. Despite these factors, however, there is a clear trend among Gartner clients to
consolidate data centers.
Network managers need to work with data center managers to design the networks in the consolidated data
centers. This may represent an opportunity to consolidate smaller Ethernet switches into larger switches, simplify
the LAN and SAN architectures, and replace less-energy-efficient switches with more-energy-efficient ones.
Back to Table of Contents

Virtualization

Virtualization pools IT resources in a way that masks the boundaries of the resources from resource users.
Overall, virtualization improves utilization, increases agility and makes I&O more manageable. Today,
virtualization efforts primarily focus on servers run at low average utilization levels (typically, 12% or less).
Virtualization can increase server utilization fourfold or more, resulting in cost structure improvements of
approximately 50%, based on our modeling. Given the benefits of server virtualization, we are projecting rapid
growth here: In 2008, only about 12% of x86 server workloads were virtualized. In 2012, that percentage will
have increased more than fourfold.
All I&O systems will be virtualized or heavily affected by virtualization. Storage has already been virtualized, but
primarily within the scope of individual vendor architectures. Virtual SANs (VSANs) are part of this trend. PCs are
being virtualized to separate the application, operating system and hardware layers, so that one can more readily
be changed with little impact on the other layers. Networking is highly virtualized with virtual LANs (VLANs) and
VPNs. We also see networking within the data center being more virtualized in another sense by having
resource pools for networking functions and routing traffic to and from these pools as needed for load balancing,
Secure Sockets Layer (SSL) encapsulation, etc..
These separate systems are aggregated and integrated via management and control systems. Service governors
are runtime control engines that take dynamic actions based on IT service demand, resource supply, service-level
agreements (SLAs), business priorities and other policies.
To accommodate server virtualization, design data center networking with high input/output (I/O) and
throughput. This will enable multiple VMs per physical server and shifting workloads from server to server.
Additionally, enable networking flexibility by introducing top-of-rack switching, implementing terabit-class
Ethernet switches, and the pooling of network resources and functions within data centers.
Back to Table of Contents

Convergence
Convergence has been a major trend in networking for more than a decade, with multiple protocols to IP, multiple
LAN technologies to Ethernet and voice onto a data platform. Although in its infancy, another convergence is
under way that involves the compute, storage and network assets of the data center. Cisco, HP and IBM have
been driving this trend, which vertically integrates server, storage and network systems and components with
element-level management software that lays the foundation to optimize shared data center resources efficiently
and dynamically. This critical connectivity layer has the potential to enable speed and agility in provisioning,
configuration and repurposing. One type of implementation of such convergence is fabric-based infrastructure.
A converged infrastructure will increase the exposure to the product integration offered by one vendor or an
alliance of third-party vendors. Initially, the convergence trend will not remove the boundaries completely. It is
likely to create new silos in the data center. In the longer term (beyond five years), some cross-infrastructure
convergence is likely to emerge.
Ethernet has essentially won the "platform wars" except for the data center. Because of many factors, such as
embedded base of alternative technologies, our view is that enterprises are unlikely to begin to deploy allEthernet solutions until 2012 at the earliest.
Network managers need to learn about the benefits and pitfalls of data center infrastructure convergence and
address this important trend by taking the following actions:

Implement top-of-rack I/O convergence solutions to simplify cabling and improve I/O flexibility.
Deploy fiber optics to all server racks and storage arrays to simplify all Ethernet solutions in the long
term.

Limit InfiniBand to applications in which its unique attributes are required.


Reorganize to include all data-networking assets under one manager to contain "platform wars."
Back to Table of Contents

New Client Architectures


In the PC world of the past 25 years, the operating system and applications resided on the desktop (some large
and complex applications, such as ERP, were located on servers that could be remote from clients). Today, things
have changed. The operating system as well as the application can be executed on the PC or a server or
streamed to a PC when needed. Choice of architecture depends on user needs and the time frame for
implementation. No longer does one size fit all.
The new choices can have substantial implications on networking. For example, with hosted virtual desktops
(HVDs), the PC operating system and all applications run on a data-center-resident server, rather than the PC
itself. Hence, an HVD architecture can require higher bandwidth requirements from users to the data center and
introduce latency issues. Software WOCs and ADCs should be considered, and they should be deployed to improve
network performance.
Back to Table of Contents

ITIL
Business has pushed IT organizations for commitments on end-to-end service quality. The Information
Technology Infrastructure Library (ITIL) is a framework (see Figure 4) that can help achieve this goal. ITIL defines
best practices in specific processes for the service desk function. It is a starting point to define your IT operations
management (ITOM) processes, but it's likely that you'll want to adapt it and revise it to fit your environment.
Figure 4. The ITIL Framework for ITOM Processes

Source: ITIL
Back to List of Figures

Back to Table of Contents

During the past decade, IT has witnessed a growing emphasis on the ITIL. Released in July 2007, ITIL version 3
uses a life cycle concept to expand and enhance version 2. In the diagram shown in Figure 4, the major elements
of ITIL version 3 are graphically depicted.
Formalizing and more systematizing NOC processes via the ITIL framework can offer the following benefits:

Reduced operational errors


Faster completion of operational processes
Greater workload segmentation by skill level
Improved service levels

Gartner advises network managers to allocate a training budget to educate appropriate networking professionals
about ITIL version 3 concepts.
Back to Table of Contents

Operations Integration
In most enterprise IT organizations, network operations are primarily organized around major network technology
platforms, such as voice telephony systems and LANs. The convergence of these platforms is a major shift that
will require changes in organizational structure.
ITIL implies an organization centered around processes that deliver IT services and this is the direction we
recommend network managers take for their NOCs. First, we advise structuring the NOC around the ITIL
processes, with technology platforms secondary to such a process orientation. To enhance end-to-end IT service
delivery, we advocate greater integration of the NOC with IT infrastructure operations. This change (see Figure 5)
will be complex and is likely to occur on a process-by-process basis. At times, the intermediary step of NOC
process structure will be leapfrogged most enterprises already have an integrated service desk, for example.
Figure 5. NOC Transition

Source: Gartner (October 2010)


Back to List of Figures
Back to Table of Contents

IT (and Network) Services


By year-end 2012, as much as 30% of large enterprises will define a core set of business-oriented IT services,
including formal SLAs. This will an increase from fewer than 15% today. However, many IT departments define IT
services as processes, such as change management. Although well-executed processes are critical to delivering IT
services, they usually do not carry great meaning to the enterprise, and it's difficult to demonstrate what
constitutes value. To get around this issue, one must think about what the enterprise uses IT for. In I&O, the
answer can often be summed up as accomplishing the following:

Running enterprise applications and associated computations that are necessary to perform the
processes of the business

Storing and backing up essential data securely, so it can be retrieved unaltered as fast as necessary
Communicating among employees, suppliers and customers to make essential business decisions and
complete business transactions

Each one of these processing categories, storing and communicating, can be further defined. For example,
communications (or connectivity) may consist of the following:

Telephony
Messaging (e-mail, instant messaging, etc.)
Conferencing (e.g., audio and video)
Collaboration
Data transfer
Presence

These categories can be provided for different service levels (such as uptime) and for different enterprise sites
(ranging from headquarters, to employees in the field or working from home). So, our key principle is to define

services in the way the enterprise uses them, not necessarily what IT delivers as processes.
Back to Table of Contents

Automation
For more than a decade, management tool vendors have promised a manager of managers (MoM) that would
automate and integrate intensive day-to-day and tactical processes. The MoM concept is taking on a new form
known as "run book automation" (RBA), named after the run books that once documented operational processes
for mainframes.
In essence, RBA provides the ability to design, build, orchestrate, administer and report workflows that support IT
operations process. A run book process can cross all management disciplines and interact with all types of IT
infrastructure elements (hardware and software). RBA products have an orchestration interface to design,
administer and monitor processes, as well as a workflow to support the processes. Integration with IT elements
and IT operational tools is needed to support processes such as change management. However, we do not see
RBA as maturing for another two years or so. Until then, there are key networking tools to add to your portfolio.
We recommend that network managers monitor the evolution of RBA to properly plan for it, as the technology
matures. Invest in management tools that lead to RBA.
Back to Table of Contents

Cloud Computing
The promise of cloud computing is ubiquitous access to a broad set of applications and services, which are
delivered over the Internet and related networks. To deliver on that promise, the cloud must provide a rich set of
network services to a broad set of applications and services.
Not all applications are the same. Some will require only the basic capabilities available on the public Internet,
while others may need an overlay on top (the "augmented Internet"), or even a private, Internet Protocol (IP)
network with application-specific capabilities. Figure 6 summarizes the characteristics of various options and
provides several case scenarios in which each might be applicable.
Figure 6. Options, Characteristics and Examples of Use for Network Cloud Computing

Source: Gartner (October 2010)


Back to List of Figures
Back to Table of Contents

Network managers should test or emulate the performance of cloud-based applications in all geographies where
they plan to deploy them. Latency can cause dramatic differences in application response time. ADCs and WOCs
may be required. In addition, consider interconnecting networks directly for example, via providers' secure
Internet gateway services between MPLS and the Internet.
Back to Table of Contents

2010 Gartner, Inc. and/or its affiliates. All rights reserved. Gartner is a registered trademark of Gartner, Inc. or
its affiliates. Reproduction and distribution of this publication in any form without prior written permission is
forbidden. The information contained herein 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's
research may discuss legal issues related to the information technology business, Gartner does not provide legal
advice or services and its research should not be construed or used as such. Gartner shall have no liability for
errors, omissions or inadequacies in the information contained herein or for interpretations thereof. The opinions
expressed herein are subject to change without notice.

You might also like