The fiscal frugality of the past few years has left its mark on IT, and nowhere is it more evident than in data center management. Companies are increasingly looking for ways to lower costs in the data center, from slowing storage acquisitions to streamlining operations.
As companies have found, the years of free spending, combined with a sprawling infrastructure from years of growth, mergers and acquisitions activity, can yield an IT infrastructure that's ripe for reduction. "Any company that's gone through an acquisition might be running three of the largest enterprise framework software," says Richard Ackerman, a senior consultant at GlassHouse Technologies Inc. in Framingham, Mass. "Reducing the costs of maintenance alone by standardizing on one can save considerable money."
In fact, doing more with less seems to be the watchword for lowering data center costs, as companies look at many ways of increasing efficiencies through consolidation, automation and standardization, according to John Sing, senior consultant at IBM Systems Group Business Continuance Strategy and Planning practice. "You need to look for redundancies, from a cost basis as well as from an IT management basis," he says. Some of the most effective ways to reduce costs follow below:
This applies in both the larger sense, as companies seek to consolidate multiple data centers down to a more manageable number, and in the smaller sense, as they also seek to reduce the physical size of the remaining data centers. "Consolidation is happening on all levels, from the data center through to the servers," says Tony Asaro, a senior analyst at Enterprise Strategy Group in Milford, Mass. A big part of the cost savings involved in shuttering entire data centers lies in staffing reductions, so data center consolidation should be planned at the highest level to create a strategy that will allow companies to continue to function smoothly during the contraction process.
Also, many chief information officers (CIO) are turning toward technologies that help maximize the usage they get out of their square footage in the data center. "The cost of expanding floor space is enormous," says Jim Damoulakis, chief technology officer at GlassHouse. "For example, a lot of customers in the financial services industry in New York City have to deal with high real estate costs, even with their data centers in outlying areas." He advises companies to look for older, bulkier equipment that has depreciated enough to jettison, and replace it with space-efficient virtualized blade server stacks. "You can get racks of Wintel servers consolidated onto one rack of blades, and it's a significant savings in terms of floor space and power," he says.
Consolidate servers and storage
As companies move to reduce the massive server sprawl found in many data centers, they apply the same logic to storage, which has moved from DAS to NAS and SAN storage. Now, many storage directors are applying high-level capacity planning strategies to organize and consolidate storage further. "The whole initiative towards server consolidation through blade and virtualization really does go hand in hand with storage consolidation initiatives," Damoulakis says.
By creating a set of metrics to measure utilization of infrastructure components such as servers, storage and networks, IT directors can compile and analyze the resultant data, find new methods to streamline operations and forecast future capacity needs more efficiently.
Some companies are even consolidating SAN islands to a larger, more monolithic SAN, Damoulakis says.
"I know one company with a very large number of independent SANs that had a port utilization count of only 26%," he says. "One direction is to consolidate into a larger SAN to better leverage that capacity." The resultant SAN will be complex and difficult to manage, so companies must have the appropriate levels of IT management policies, procedures and controls in place to handle the new entity.
Centralize, standardize and automate
As part of an overall data center management effort, senior-level IT executives need to choose a standard list of vendors, and enforce adherence to that list. Supporting fewer technologies allows IT staff to provide better support, and it also lets the CIO take advantage of economies of scale and set up preferential purchasing with those vendors.
Storage managers should also research how automating and centralizing storage management processes can help them reduce costs. While some of those savings may come through staff reduction, by and large, the management tools will let existing staff do more work more efficiently. "Automation is important," Sing says. "It keeps staffing resources flat."
"You can gain efficiency by devising really strong processes and procedures, and then automating and maintaining them centrally," Ackerman says. Functions such as reporting and monitoring, which allow storage administrators to monitor, provision and support storage centrally, are especially attractive.
Take on disaster recovery
Companies that run more than one data center can also save money by backing up one site to another for disaster recovery, but this strategy bears a significant amount of risk analysis. "A large number of customers that we speak to are interested in implementing some such form of disaster recovery," Asaro says, "but there's a perception of a high cost associated with it."
The potential for cost savings can go either way, agrees Damoulakis, but still, "It's done fairly commonly, and we know of a number of companies that have moved from a third-party site to their own disaster recovery environment," he says.