The Small Revolution Of The Data CenterFinancial services firms are rethinking how--and why--they build data centers, turning traditional data center models on their heads.
For firms like JP Morgan, that has meant turning old computing models upside-down. "In the legacy model, you used to build an application and you would create the infrastructure to accommodate it," says Peter Ahrens, managing director and chief technology officer at New York-based JP Morgan. "So the applications themselves created the infrastructure design."
- Data Center Performance: Optimization Secrets Revealed
- Growing Your Datacenter Without Growing Pains
- IDC Analyst Connection: Using Blade Systems to Cut Costs and Sharpen Efficiencies
- Data center consolidation restructures your IT costs for continued growth: New discovery tools determine logical and physical move dependencies to help limit risk
In the new model, the bank builds the infrastructure, and the application is designed to fit the existing environment, Ahrens explains. "The application is less customized, but it's available on demand. And it's cheaper," he adds.
One of the biggest challenges firms traditionally face when they plan a new multimillion dollar data center is that the facility often represents a 10-year investment, and they must try to envision their power, server and storage needs for the next decade before they start building. But with technologies and storage and data requirements evolving so rapidly, current needs may differ wildly from what a firm will require in just a few years.
"We would historically have built or provisioned data centers on the basis of things we knew at one point in time," Ahrens relates, adding that engineers would have planned for a specific amount of rack space, storage and cooling, and electric distribution would have been consistent with what they knew at that time. "But today," he says, "some dynamics are changing."
And they are changing more and more rapidly. For example, the overall mix between the number of servers and storage capacity has shifted, Ahrens points out. The power of servers today means firms can run fewer in the data center than historically, he says. Meanwhile, use of storage has been multiplying. "Five to eight years ago, you would have had a small number of storage devices and a large number of server racks," he comments. "Today the trend is toward more space consumed by storage and less by server racks."
Other traditional assumptions about the data center that have been turned on their heads include power requirements. "Some new technologies require far more power," Ahrens says. "Even if you have room on the floor, you may [no longer] have adequate power or cooling to run devices in that location."
While JP Morgan, like most major Wall Street firms, would once have built large facilities according to its expected future requirements, it is now building more-flexible data centers in what Ahrens calls a "pod-like" fashion. "In a couple of recent acquisitions, we built a large facility and put a large shell over it," he says, explaining that the bank builds out one pod, or room, in the facility at a time. "We're trying to match up our investment stream more closely to when we need it--in other words, to only spend money when we need it.
"Our business drivers are, 'Don't spend money and build stuff until you need it,'" he continues. "And by doing that, you are better able to build the pods to the specs of what you require." For example, as JP Morgan builds out a current pod, the firm's engineers might focus on the fact that power density and floor loading will need to be designed more for storage needs rather than server requirements, Ahrens explains.
In addition, like other capital markets firms, JP Morgan has been busily consolidating its existing facilities. "We are reducing the number of data centers we have. And we will be building larger facilities with more expansion capability, rather than having many facilities scattered around the place," Ahrens reports.
IT is caught in a squeeze between requests for new applications, services, and device support and demands from upper management to keep budgets lean, staffing light, and operations tight. These are irreconcilable objectives as long as we spend the vast majority of our resources on legacy services. Read our report now. (Free registration required.)