October 03, 2012

The numbers are in, and no matter how you slice and dice the metrics, they don't look good. Whether it's the declining revenues of capital markets firms, the cost of energy or the underutilization of servers, the numbers do not paint a pretty picture for the future of the data center.

While there always will be a place for data centers in the financial services industry, the role of the traditional corporate data center is changing. Firms still can gain certain competitive advantages from maintaining and running their own data centers, such as latency reductions. And, in an industry that has collectively been under regulatory scrutiny for decades, firms still require top-flight security on customer and corporate data. (As the saying goes: You can't outsource security, risk and compliance oversight.)

But the cost to run a proprietary data center just to maintain security and competitive advantages for certain applications will soon become too much to bear -- if it hasn't become so already.

For starters, simply powering a data center filled with servers humming along 24 hours a day is extremely costly. A typical corporate data center consumes millions of watts of power, even if it isn't actually processing any data.

In fact, despite years of trying to get firms to increase their utilization of existing servers, most servers in data centers run at only 7 percent to 12 percent of capacity, according to separate studies from Gartner and McKinsey & Company, as reported in The New York Times. This means that most of the energy consumed -- between 88 percent and 97 percent -- is used to do things other than actually process data, such as cooling.

[The Death of the Data Center Has Been Greatly Exaggerated ]

Meanwhile, financial firms are pushing to reduce costs across all business units. It's no secret that technology budgets have been squeezed in recent years. But while most data centers are not being used to capacity, the costs to run them remain relatively fixed. And often data center executives are afraid to shut down underutilized servers for fear of losing their jobs -- just in case something goes wrong.

At the same time, many applications have moved to hosted offerings, either by software vendors or to purely cloud-based solutions, such as Salesforce.com. This leaves data center executives in a pickle: Where do the costs of running the data center get allocated? Most companies allocate data center costs to the business users who consume computing power. But if fewer users are actually running applications in the data center and the costs remain the same, the costs for the remaining users will increase. (Try telling that to your business partner: You will get the same service as last year, but it will cost you 20 percent more.)

This doesn't mean that data centers in financial services will disappear today, tomorrow or even by 2018. But unless firms figure out a way to control data center energy consumption, increase server utilization, and provide the features and flexibility that are driving many business users to the cloud in the first place, data centers as we know them today on Wall Street will shrink drastically over the next few years. It's simple economics: The numbers don't support the status quo.

ABOUT THE AUTHOR
Greg MacSweeney is editorial director of InformationWeek Financial Services, whose brands include Wall Street & Technology, Bank Systems & Technology, Advanced Trading, and Insurance & Technology.