Although the notion of consolidation is not new, advances in open system
functionality on the part of hardware vendors over the past few years
have paved the way for increasing opportunities to consolidate. The linked
goals of reducing IT complexity, lowering total cost of ownership (TCO),
and providing a better way to access and manage an organization’s information
and resources are inherent to server consolidation initiatives – and are
now more attainable than ever.

For many organizations, consolidation is not just a tactical initiative
but also part of a larger, strategic data center direction. So companies
are using consolidation to control costs and complexity and also combining
consolidation with other infrastructure initiatives to move toward a more
flexible and responsive overall architecture to meet the challenge of
today’s demanding business environment.

The Road To Server Consolidation

There are several different ways to consolidate, from simply bringing
existing systems under common management practice to consolidating many
heterogeneous applications and services on one or a few systems. But all
server consolidation initiatives involve several steps, from high-level
consideration of business goals to tactical planning and timelines. Since
consolidation can be quite complex, many companies start with a pilot
project and then leverage their experience into larger efforts. Some IT
organizations have the resources to perform consolidations in-house, but
most depend on outside service firms for at least portions of the project.

TCO and More

For many companies, the overriding benefit of a server consolidation
project is more-efficient use of resources and lower TCO. New server technologies,
such as domains and dynamic reconfiguration, enable IT managers to allocate
resources dynamically and manage service levels, thereby enabling better
utilization and control of system resources.

By integrating a server consolidation initiative with flexible systems,
IT management can not only reduce the number of hardware and software
platforms it has to deal with but can also apply more standardized procedures
and disciplines to a streamlined and often recentralized environment.
Simplified management leads to improved service levels and uptime for
the organization, which results in more efficient resource utilization.

Although underutilization is common in distributed environments, many
companies also have systems that are taxed to the point that workloads
exceed capacity, leading to degraded response time and even system downtime.
In cases where technicians develop, test, and run production environments
on the same system, these groups can work at cross-purposes, with developers
testing applications by pushing them to the limit while production managers
strive to keep the environment stable.

Server consolidation enables an organization to plan how applications
are combined onto specific systems and to apply consistent management
practices that keep systems and information available. A more effective
computing environment then leads to better skills utilization. With distributed
systems, many users have had to perform double duty as systems administrators
– a job they’re often not prepared to handle effectively and rarely learn
thoroughly. When systems are consolidated and recentralized, an experienced
data center professional can do a much better job of bringing together
multiple disparate platforms and running them as a single seamless environment.

Companies that have consolidated require fewer administrators and managers
and can often re-deploy them from routine maintenance tasks to strategic
development projects. At the same time, a company that is consolidating
its systems has a better chance of justifying the hiring of an expert
staff. A consolidated environment also facilitates proactive, rather than
reactive, management.

Figure 1
Client-server trends

Figure 1

Planning a Strategy

The steps involved in a server consolidation strategy vary from company
to company, but most would agree on the following general outline:

Business Goal Clarification

First, companies need to prioritize their business goals and decide what
they want consolidation to do for them. These goals might include reducing
total cost of ownership, deploying new or modified applications faster
and more effectively, providing accurate and timely information to improve
decision-making, improving user satisfaction, providing better service
at lower cost, and many other objectives that all leading organizations
are striving to reach.

Once business goals are defined, support for technical goals can be determined.
For example, to reduce TCO, companies may want to reduce storage costs
associated with data replication, purchase fewer large systems, and reduce
third-party software costs by supporting fewer versions.

Asset Identification

With the goals clear, the first tactical step in server consolidation
is to inventory the environment. This step can be difficult and time-consuming,
but to achieve the best-possible return on investment, it is critical
to know exactly what hardware – everything from memory modules to storage
devices to servers – is deployed in each consolidation target area as
well as what operating systems and revision levels, databases, custom
software, and packaged applications are in use.

Capacity Study

With resources inventoried, the time has come to determine the right
types and number of servers for a pilot project. The best targets for
consolidation are systems with stable configurations and well-understood
usage and performance characteristics. It’s also important to understand
the operational requirements of the consolidation targets. For example,
high priority should be given to evaluating issues such as downtime, batch
job windows, disaster recovery scenarios, backups, concurrent users, and
response time for each of the systems that will be consolidated.

Proposal and Plan Development

The consolidation proposal and plan document should take the information
gathered from the asset identification and capacity study and recommend
a configuration for the consolidation server. Some proposals, for example,
include a TCO study, developed from a database of acquisition costs, depreciation,
operating costs, and other customer-supplied information that provides
a basis for comparing distinct solutions. In addition, the proposal and
plan should give a detailed timeline for implementation.

Consolidation

Some companies have the in-house expertise to carry out a server consolidation
pilot project from start to finish. Many, if not most, companies, however,
seek out a consulting firm with expertise in enterprise servers and storage,
system software, networking, and database environments. If porting or
special customization is needed, consultants often enlist the support
of outside technical resources.

In addition, a consulting firm should be able to help design reliability
and performance tests for consolidated systems; recommend or offer software
development and system management tools; keep technical personnel current
on consolidation trends; and keep management apprised of the results that
other, similar consolidation customers are experiencing.

Choosing a Platform

In evaluating hardware partners and platforms for a consolidation project,
it is important to take a look at the kinds of technology they offer,
to see whether those technologies mesh with the company’s vision of what
consolidation should ultimately do. Because consolidation means moving
various applications and services onto fewer machines, the platform being
considered should provide strong solutions in the following areas:

Scalability

To have the flexibility to add users, data, and applications easily,
the consolidation platform should be scalable. IS managers with scalable
platforms under their management are able to add processing power, I/O,
memory, and bandwidth, quickly and simply. The server platform should
run a single operating system to reduce application porting problems,
simplify network and system management, and allow the topology to change
quickly as the business changes and grows.

Availability

The server platform should not only provide exceptionally high performance
but also deliver on expectations for uptime and application availability.
With more applications and services residing on fewer systems, it is more
critical than ever that downtime, both planned and unplanned, be kept
to an absolute minimum.

It is necessary to specify systems that have no single point of failure
and to configure redundancy in without adding complexity and cost. The
systems should also enable companies to make changes on the fly – to change
system configurations while keeping applications up and running. Consider
systems that utilize domains, to keep problems in one area of the server
from affecting a mission-critical application in another.

Resource Management

The ability to manage resources effectively is the key to getting the
most out of the consolidation effort. The ideal consolidation platform
allows system administrators to create, manage, and bind processes and
applications onto processor sets within a domain or system. A fair-share
scheduler should be included in the operating environment, to ensure that
each application receives the appropriate amount of system resources.
Companies should seek technology that allows borrowing of resources from
elsewhere on the system when needed and enables management from anywhere
on the network.

The ultimate goal of many companies is to consolidate on a large scale,
perhaps across the enterprise. In most cases, this means some consolidation
projects at the workgroup level, some on midrange servers, and some at
the enterprise level. So working with the vendor with the broadest binary-compatible
product line is essential. The product ine must support interoperability
with Windows NT at the low end and provide the highest levels of flexibility
– along with the technical capabilities discussed previously – at the
mid-range and high end. The vendor should also provide a scalable, robust
operating environment that can support consolidation at all levels.

A critical requirement is to evaluate potential consolidation partners
not only by the technologies and products they have to offer but by the
overall commitment they make to customers as well. Does the company offer
an integrated product line, without hardware and software discontinuities?
Is the vendor committed to protecting its customers’ investment by ensuring
binary compatibility now and in the future? Are attractive trade-in programs
available?

The Politics of Server Consolidation

In today’s global business environment, change is inevitable and ubiquitous,
and server consolidation is about fundamentally changing the way IT is
viewed and managed in a business. The technical issues, although complex,
can often be resolved with a minimum amount of disruption in the day-to-day
workings of an organization. However, the political issues frequently
lead to infighting that can delay and sometimes derail the project. In
light of this, high-level management should become involved in and demonstrate
visible support for the consolidation process from the start.

Many companies report that once the consolidation effort is complete
and users across the enterprise start enjoying these benefits, relationships
between IS and business units actually improve. But to get to that happy
point, the consolidation-team champions need to take responsibility for
positioning the benefits of server consolidation for each group. By keeping
everyone – IS, department heads, rank-and-file users – involved in the
process, companies reap the benefits of wide-ranging input and across-the-board
buy-in. That can go a long way toward making any server consolidation
project a successful one.