Are economic conditions changing how you manage server turnover?

Guest post: TechRepublic's Scott Lowe, Chief Information Officer for Westminster College in Fulton, Missouri, examines some of the reasons server sales have dropped and handicaps the recession's impact. For more posts like this see TechRepublic's data center blog.

Between cloud computing, virtualization and economic conditions, the data center has certainly changed form over the past ten years. Where there used to be a box for each discrete workload we now have boxes running virtualized server instances for dozens of workloads and some services run “in the cloud” and, relying on no local servers at all, simply depend on the corporate router to achieve their aims. Every day, more and more software-as-a-service vendors pop up offering their wares. And today, unprecedented economic conditions are forcing organizations of all types to deeply examine everything they do to make sure that every dollar spent directly supports the bottom line.

With this perfect storm of activity, what’s happening in the data center? In May, CNet quoted an IDC report indicating that worldwide server sales were down 25% in the first quarter when compared to sales of a year ago. In February ZDNet’s Larry Dignan quoted another IDC report indicating that year-over-year server sales fell 12 percent. There are also published reports claiming that 2009 server sales will plummet more than 20 percent for the year.

I’ve posited some possible reasons above for this downturn, but let me expand on them:

• Virtualization. With dozens of workloads now running on a single box, physical server sprawl is a thing of the past (of course, virtual server sprawl is now here to stay!). Obviously, fewer servers in the data center directly equates to fewer server sales for each vendor. Obviously, virtualization has other benefits beyond simple consolidation; for example, virtualization generally reduces service deployment time and when deployed in the right way, virtualization can be a boon for high availability.
• Software-as-a-service. I doubt that we’ve yet seen the full impact of SaaS, but I can’t imagine that it hasn’t yet had at least a minor impact on server purchases and sales. Many colleges and universities, for example, are outsourcing student and, sometimes, faculty/staff email to the likes of Google and Microsoft, thus eliminating the need for a server infrastructure supporting those outsourced email services. No longer are those old servers on the replacement cycle.
• The economy. I’ll admit that I don’t have hard data on this, but I also indicated that this is a posited list, not an authoritative one. That said, I’ve heard that server replacement cycles are being extended in many organizations. After all, even a three year old server still has reasonable life left in it, especially these days. If the server is out of warranty, there are plenty of nationwide servicing groups willing to cover the unit for much less cost than it would take to buy or even lease a new system. With budgets tight and layoffs happening everywhere, spending on infrastructure that can reasonably be extended is probably not high on the priority list for many companies. All it takes is a chunk of companies that were once on a three year server replacement cycle to extend that cycle to four or even five years and the server market takes a downward turn. Beyond replacement, big projects are being put on hold while companies regroup, thus eliminating the need for supporting project infrastructure and, hence, also nailing server sales.

Really, the items above, with the exception of the economy, are more than likely a part of overall IT efficiency and cost savings strategies that would probably take place anyway. However, with the economic situation thrown into the mix, I imagine that any possible savings measures are being exercised and organizations are doing everything possible to defer purchases as long as possible.

I can tell you that we at Westminster College are aggressively pursuing our virtualization strategy as a way to reduce overall infrastructure costs and to reduce the number of physical servers that we manage. As stated above, we’re also well aware of the ancillary benefits inherent in a well-designed virtual infrastructure and intend to leverage those benefits as well. As for new servers, we are scrutinizing all of our purchases like never before and squeezing as much life as responsibly possible out of both our desktops and servers.

Now, it’s your turn: What, if anything, are you doing differently now than you were before with regard to your server and data center infrastructure? Are you going full steam ahead with virtualization? Are you investigating SaaS opportunities? If you are looking at SaaS, what is your primary driver - infrastructure savings, personnel savings, or better service delivery? Is your organization holding the line on expenditures and expecting the IT department to extend the server lifecycle?