I. The Problem
At a time when governments at all levels are being called upon to once again, “do more with less,” Data Center Consolidation (DCC) has become a hot topic. While the Federal government has called for a reduction of 2,400 Data Centers down to about 1,200 by 2015, what does that mean, and most important, how does government get there?
Studies have shown simply reducing “bricks and mortar” data center facilities do not yield the cost savings predicted by planners as the bulk of expenses is often accrued by the Maintenance payments for the software running within the building. A reduction in property, plant and equipment expenses will be realized, but is comparatively small when measured against the cost of intellectual property being maintained within the buildings.
Another challenge confronting both business and IT leaders is the number of machines, applications and data bases running in support of an agency – often without a clear architecture or data governance plan – which makes determining which systems to retire or combine a complex problem. The business of government is unique in that many agencies have been forced to simply add resources to their environment, integrating on the fly, in order to meet changing Federal or legislative requirements often under an onerous financial or court-ordered deadline. Another issue is lack of standards, for example, unlike the Commercial market there are no GAAP principals that serve as a foundation for development of Public Sector financial systems. These might be built according to the needs of an Agency, State, City or Locale as dictated by the statute which varies by locale.
These systems have usually been implemented over years by third party consultants who then manage them for a fixed period, making changes which might be partially documented (if at all), and since these contracts run for a specific period, some large systems (ERP, Logistics, Medicaid) may have been managed and changed by a series of consultants over a sequence of contracts spanning years. Turnover in agency staff only leads to further creation of an impenetrable Gordian knot standing squarely in front of today’s policy makers and IT professionals now charged with, “Data Center Consolidation”.
Changing technology offers a variety of approaches to determining what needs to be consolidated, in what order, with an eye towards determining what data will be necessary, in what format and how frequently it must be accessed. Such concepts as the Cloud and big data are offered by industry as potential solutions, but they take on a far different meaning within Public Sector versus a Commercial environment as the underlying business process for government may vary by agency, mission or level of government. By contrast the principal behind moving data to the Cloud to gain leverage in global buying decisions is not going to vary widely between process versus discrete manufacturing industries, consumer products etc.
In contemplating a Cloud strategy, in addition to a consideration of the technology being used, public sector executives must balance concerns over data privacy, security, retention and access which are rarely factored in to Commercial re-platforming strategies. Fortunately like the shifts which came before it from mainframe to client server, and client server to the web, the drive to Data Center Consolidation can be enabled by the move to the Cloud. But that is only one approach successful agencies can deploy.
II. Approaches to the Problem
DCC can be affected in many ways; it:
- Can mean going from five data centers to two, reducing the bricks and mortar required to support a business mission;
- Can involve combining Programs and funding streams, collapsing the requisite underlying architectures needed to support the two;
- Can involve moving data and/or processing to the cloud, assuming security risks are addressed;
- May mean simply archiving old data, cancelling legacy applications, and pulling the requisite data forward to operate new applications, databases, data warehouses or business analytic tools, and most likely
- May involve some combination from all of the above.
A detailed analysis of each is beyond the scope of this piece, but successful customers have followed a similar path to achieving success.
- Ensure executive sponsorship – DCC is not an IT problem. It is a business strategy which requires executive buy in to the project, an understanding of the goals, agreement on alignment of processes and a clear vision of the end point. This may require executives from various departments or agency to coordinate the vision and project governance up front. None of this is a function of the technology.
- Build the Strategy First – Once the strategy is achieved – and although IT should not be the driver, senior IT Staff should be included in the discussion – then the selection of technologies should be made early in the project. Ironically, many of the large existing vendors have little reason to support a smaller foot print in customer data centers, so many analysts are recommending that customers and systems integrators look to tier-two vendors like Informatica. The broad portfolio of products combined with financial stability a vision for supporting data management from the main frame to the cloud offers a potent, cost-effective platform choice for public sector customers and partners.
- Assign your best people – Like an ERP implementation, or any large IT undertaking, a DCC project can empower a more nimble, cost effective enterprise able to address government’s challenges, or it can drag an agency down in to a tough, IT slog. Establish the strategy and governance, select the best technology, then give it to your best people to accomplish.
- Ensure ongoing alignment with program, technology, vendor and consulting staff. Your project work plan should have a role for each, and if it doesn’t, find one. Vendors will generally react in support of customers that encounter problems. This is made easier if each firm has some footprint early in the project so relationships are established, and problems do not require a project hold while a new employee or team is brought up to speed.
- Buy Too Much Training – Invariably one of the greatest mistakes public sector leaders make is failing to ensure employees have the skill to manage the new environment. Vendors tend not to push training, customers cut back to “save money.” Both are wrong. Training your employees is the greatest predictor of enterprise IT program success.
Data Center Consolidation is a project like many other enterprise IT projects, but one that promises an extraordinary return on investment if done well. A comprehensive vision, combined with solid project management and the right technology will afford every government customer the benefits of what started out as a “mandated reduction.”