Data center consolidation: (Past) time to get moving

Citing possible operational savings of 30 to 50 percent, the Office of Management and Budget in October 1995 issued a bulletin directing agencies to inventory existing data centers and to produce by September 1996 a consolidation plan, with work due to be completed in early 1998.

Well, early l998 has arrived. Agencies did send plans to OMB, but not much consolidation has happened. OMB and agencies were understandably distracted by the implementation of the Clinger-Cohen Act and the Year 2000 date-change problem. You sometimes hear that the spread of client/server architectures makes data center consolidation a less important issue. That may be, but agencies aren't closing data centers at this point despite the spread of client/server technology. And with huge federal databases in many agencies, data centers are likely to remain an important part of federal IT life.

Data center consolidation remains a target of opportunity for cost savings. Achieving that savings doesn't require rocket science, dramatic new ways of thinking or bleeding-edge technology. Over the past year, the General Services Administration's Virtual Data Center contract, run by Ron Decker, director of the Federal Computer Acquisition Center, and his boss, Charlie Self, who heads the Information Technology Integration Office, has been available to help agencies consolidate operations.

Because it is a multiple-award, task-order contract operating under the same procurement rules as other governmentwide acquisition contracts (GWACs), the Virtual Data Center allows agencies to contract with one of the three primes in a matter of weeks rather than the year or two this would have taken before procurement reform.

GSA also provides technical support in running the competition among the three vendors and in writing performance-based statements of work, including service-level agreements defining the performance levels expected for the data center. Self runs a customer-oriented, pro-taxpayer shop that is a model for how GWACs should be run throughout the government.

Yet the Virtual Data Center contract is begging for business. In a year, it has negotiated only one task order: a consolidation of data center contracts at the Education Department. (Hats off to Education in general and to Jerry Russomano in the Office of Postsecondary Education in particular for being out front on this one.) Four more data center consolidation task orders are in the works at GSA, of which one involves a major federal agency.

The problem is not the idea itself or the lack of good contracting vehicles but rather turf and job protection— the banes of organizational existence in general and federal organizational existence in particular. Consolidating data centers threatens the turf of those managing existing centers, and it threatens jobs. Revealingly, the Education consolidation is occurring in an environment where the agency's existing data centers are all already outsourced, so there are no job issues for federal employees. And according to folks with the Virtual Data Center contract, the agencies with whom it is now talking don't want to go public with their intentions until they have resolved touchy internal issues.

I yield to no one in my admiration for the federal work force. And it's natural for people to be concerned about job security. But at some point enough just has to be enough. Data center consolidation is such a low-hanging fruit for significant dollar savings— and performance improvements, through the use of service-level agreements— that we are going to need to tell those who might be affected that they are going to have to look elsewhere. It's hard to imagine, with a booming economy and a shortage of workers with IT experience, that this will be a big problem. If we're going to make this necessary transition, a better time than now could hardly be imagined.

The data center issue is about to get hot again. OMB has rewarded those few agencies that have worked to consolidate their data centers based on its l995 policy by allowing them to keep budget savings they achieve through consolidation. However, the president's fiscal 1999 budget is expected to reflect savings— expressed as a certain number of dollars for every million instructions per second of processing power— that will be deducted from agency budgets, based on the assumption that centers will be consolidated. (The number is conservative enough so that most agencies will achieve larger savings if they do consolidate, which will allow them to reap some dollar gains from consolidation despite the new policy.) Agencies that don't consolidate will take a budget hit, which will affect programs and workers elsewhere.

The Education consolidation is scheduled to be up and running sometime this month, and with any luck, it will show savings that will be hard to ignore. Stay tuned.

-- Kelman was the administrator of the Office of Federal Procurement Policy from 1993 to 1997. He is now Weatherhead Professor of Public Management at Harvard's Kennedy School of Government.