Server Consolidation: Steps to IT and Business Rationalization

June 18, 2003 (Computerworld) In several past columns, I have mentioned the possibility of server consolidation. Whether replacing or adding an operating system, companies have opportunities to consolidate servers to reduce equipment -- and more. In the past five years, organizations have been adding servers and storage at an incredible rate. There are several reasons for this:

• Customers purchase most servers to support a single application.


• They perform limited capacity planning and management.


• When they run out of room to add storage, companies buy another server.

A company that starts with 100 servers and has a history of 15% annual growth will have 201 servers in five years' time. Must that rate of equipment purchase be sustained to support business growth? Could annual investment at just 5% ensure sufficient capacity?

	Christopher Burry is technology infrastructure practice director and a fellow at Avanade Inc., an integrator for Microsoft technology that's a joint venture between Accenture Ltd. and Microsoft Corp. Readers are invited to send comments or questions to Burry at Christopher.Burry@avanade.com.

This compounding effect is present in storage as well. Despite typical targets of 50% utilization for direct-attached storage and 80% for network-attached storage, I have found that a 25% usage rate is fairly common.
Once IT managers decide to consolidate, they tend to focus on replacing function. But server consolidation impacts operations, applications, data and security. More than reduction for reduction's sake, the real value comes from rationalization of IT resources, freeing them for projects and tasks with greater value.
Having worked both with customers who have highly centralized operations and with customers who have a server in every office, under every desk and in every broom closet, it has become clear that the most effective model is somewhere in between these extremes. It doesn't make much sense to return to an extremely centralized environment just to reduce the number of devices in use. Experience suggests that not everything that can be consolidated should be.
The first step in developing a server consolidation strategy is to identify the systems, processes and people that could be combined. An organization might consolidate support staff and processes, relocate distributed systems, or both -- reducing support and computing resources to fewer systems and sites.
The next step is to identify the wider impact on operations. Specifically, I recommend that IT executives pinpoint the following:

• Data that could be centralized.


• Similar applications that could be centralized.


• Different applications that could be consolidated.

It's also wise to consider not only performance challenges, but also business problems. This exercise can reveal processes that will be affected -- and could be remedied -- by consolidation. For example, support calls may have increased, not because of platform problems, but because the sales force is using a database as a CRM tool. How can consolidation help to change -- and prevent -- this behavior?
In working with a large financial services customer, this kind of assessment revealed that opportunities to reduce equipment were inherently linked with process changes. With 1,000 application servers and 650 preproduction servers, it seemed quite clear that there could be a chance to eliminate equipment. However, further examination determined that a lack of a standardized testing strategy had led to the proliferation of preproduction hardware.
Likewise, the number of infrastructure servers (700) seemed excessive. However, an understanding of the business proved crucial. Operating many branch offices, this institution would not have benefited from centralization of infrastructure, and we reaffirmed that there were benefits to having a server in each branch office, including gains in security.
Finally, while consolidation can dramatically slow annual hardware expenditures, reduce the cost of operations and improve organizational resiliency, it may not solve all technical -- or business -- headaches. Therefore, I advise IT pros to manage expectations and build measurement into their plans. Tangible and intangible benefits include the following:

• Software and hardware cost reductions, through the termination of hardware leases and the resale of equipment, as well as reductions in server licenses, redundant products and support.


• Simplified security requirements.


• Simpler integration of new technology.


• Centralization of management and maintenance.


• Better utilization of staff skills across a unified environment.


• Consistent, responsive service levels and quality assurance as a result of standardization of processes.

There are plenty of reasons for IT executives to plan and measure consolidation's impact on technical and business performance. In February, when it announced its "10 Step Guide to Achieving Business Value of IT," Gartner Inc. estimated that more than 60% of all IT projects are initiated by business units. If IT departments weren't already convinced of the need to factor business operations and goals into consolidation strategy, they should find this impetus to do so.