Does a diverse IT infrastructure save money?

A recent survey executed by MeriTalk, sponsored by Brocade, promotes the notion that adding more suppliers can reduce IT infrastructure capital costs and only minimally increase operational costs in Federal data centers. Does this survey provide useful information, or is it yet another self-serving, badly executed study?

From time to time, I find my inbox full of self-serving, badly executed, limited studies presenting themselves as scientific, balanced sources of useful information. This time I've received breathless notices of at least ten different studies that purport to present what IT decision makers are thinking. One of the more useful ones was executed by Meritalk and sponsored by Brocade. The study focuses on creating a more diverse IT infrastructure that can reduce capital costs while only minimally increasing operational costs in U.S. Government installations. Two hundred and two IT decision makers in U.S. Government installations were asked about the costs associated with the infrastructure of their data centers.

Here's a quick review of the study findings:

  • 95% of respondents believe that there are benefits to using more than one manufacturer in an area of their Agency's IT infrastructure
  • 45% believe that adding a supplier can drive down costs — a reduction of as much as 20% of their annual budget.
  • 5% of the respondents say that their data center is using a single supplier. Another 23% say that they use just two or three suppliers.
  • 41% state that their agency has not even considered introducing additional manufacturers
  • 76% of the respondents indicate that their Agency specifies a supplier at least some of the time. 41% state this is true all of the time.

Snapshot analysis

Although it is clear that the study was designed to support Brocade's attempts to become a major portion of the U.S. Government IT infrastructure and reduce the penetration of both Cisco and Juniper, the analysis appears relevant to areas of the IT infrastructure other than application software. It also appears likely that the results of the study would be useful to large, non-government IT installations.

Rather than focusing on an architecture, a structured plan on functions that need to be accomplished in any given area of the IT infrastructure, organizations often take what they believe is a simpler approach: to standardize on products. An example would be to standardize on Windows as an operating system rather than specifying that the operating system should run on a specific type of hardware, offering applications that support certain functions and can be managed by a specific set of management tools. Another example would be to select a single supplier of systems, such as Dell, IBM or HP, and then standardizing on a set of system capabilities.

This approach locks the organization into a single vendor's products and creates a rather inflexible and, if the study is to be believed, more costly environment.

The study supports the notion that adding more competitors to the list can drive down capital costs and, if done properly, can manage the increase in operational costs that come from the requirement to manage multiple suppliers, multiple management tools, and other operations.

If you'd like to review the study yourself, it can be found here.