Banks are struggling to meet the storage, power and cooling requirements of newer servers that crunch customer and operational metrics, says Aperture Technologies CEO William Clifford, former Gartner CEO and ex-ADP executive. In financial services, storage needs are growing at a pace of 25 to 30 percent.
Why do banks need enterprise infrastructure management of data centers?
Banks have all the same issues and suffer all the same problems that other enterprises do, but at heightened levels of needs. On average across all industries, server growth may be expanding by 10 percent; within the financial services industry, it's probably 25 percent to 30 percent growth of storage. It has been growing at exponential rates through the '90s, and exploded in the e-commerce era.
And banks have trouble keeping up?
In the days of the mainframe and even client servers, it was much easier to do these things. Data centers were built with a 10 to 15 year half-life. They were built, they were powered, they were provisioned with cooling, raised floor tiles were installed and that was about the last time you ever saw the facilities people in the data center. The IT guys then took over, and dropped mainframes and boxes on the floor. That's not possible any more. [Looking back], I realized how much I didn't know about operations of the 43 ADP data centers I was so intimately involved with. I didn't know all of the aspects that lead to the capacity management of the data center. I didn't understand the detail within the power and cooling requirements in each of the devices and how critical of the reliability of the data center was to all of these issues. And I was titularly the executive in charge.
Why do old data center organizational models hinder banks?
You look at a data center that has 70 percent-plus of the physical square footage empty, but [can] be out of capacity because capacity is no longer simply a function of the size and square footage of the data center. You've got [minimal component] blade servers being delivered, installed and packed efficiently from the point of view of square footage, but totally inefficiently in terms of the heating and cooling capabilities and power distribution capabilities of data centers built as recently as five year ago.
Your firm claims that many organizations have a "chaotic level" of maturity for data center infrastructure. What are some examples?
When we attempt to go and work with a client and load our repository, we get some Excel spreadsheets, or we get a link to some procurement or asset management database that has the stuff bought in the last 12 months. Somebody comes in with a napkin, literally. They have no idea what's on the data center floor. They have no idea what applications are running. Even the most sophisticated organizations, because they were relying on unenforceable manual processes, have no escalation, no workflow management, and no business rules that had been written. The data center was typically managed in a very ad hoc way.
Why do basic things like space, cooling capacity, and power management get lost or misapplied in operational capacity plans?
These are generally being extrapolated. We had a client who came in and was talking about capacity management. They got a little bit anxious that the floor was being filled, but the way that they did capacity management was they took the historical growth in the data center square footage, and said, every year, we seem to be using more and more square footage, so we're going to need a new data center...without understanding that's only one of about four of the drivers.
Since your days at Gartner, what impact has technology maturity made on this approach to enterprise data center management?
It's just an indication of professionalization of management. The need to manage the data center as if it was another key critical operational entity within the enterprise, and not a one-off, blue-collar facility where they print the check statements. What's happened across the industry, managers are now IT savvy, they look for a dashboard and they are looking for metrics.
