How important are blade servers to adopting virtualisation?
Or more accurately, to what extent does the choice of server vendor play a part in the success or failure of a virtualisation rollout
That was the theme of an interesting discussion over lunch with Dell's VP for software and solutions enterprise product group, Rick Becker. To no great surprise, Becker thinks the choice of vendor is extremely important, and to even less surprise he thought that Dell had it wrapped up. " We work closely with Vmware in optimising our blade servers for virtualisation," he says.
It must be particularly galling then when HP claims a new benchmark record for running VMware but Becker is blasé about his rivals' achievements, "It's only a few hundredths of a second," he says sniffily, rather spoiling the studied indifference by noting that Dell will be looking at ways of snatching the record back.
But beyond the arcane world of benchmarking, a world where vendors argue over fraction of seconds, there is a wider issue. Much of the debate on virtualisation has centred, quite rightly, on the underlying software but we at Techworld know - because we examine the search terms that are being used on the site, and "what blade server for virtualisation" is an extremely popular question - that users are equally concerned about the hardware platform and which blade server should be adopted.
I'm not convinced that the sort of benchmarking carried out by test labs is particularly helpful within the enterprise. They're great for giving a guidance and they certainly excite marketing professionals from the vendors but they bear little resemblance to what real-world IT professionals are looking for - the savings of a few hundredths of a second under test conditions matter not when it comes to keeping mission-critical servers up and running.
What then decides the choice of server? Is resilience the most vital aspect? Is speed important or is cost much more crucial? How about the quality of technical support? Is it ease of handling? Is it connected with the company's desktop policy in any way? Do customers look at power consumption and the carbon footprint or is that so much greenwash?
We're currently in the process of setting up a series of tests to examine how all the major blade server vendors handle virtualisation and we'd love to hear what issues you'd like to see examined and what's important to you.