A Microsoft employee asked me for more information on the recent Virtual Reality Check performance analysis comparing VMware vSphere 4.0 Update 1, Citrix XenServer 5.5 and Microsoft Windows Server 2008 R2 Hyper-V running against the new workload simulator, Virtual Session Indexer. VSI is provided by Ruben Spruijt and Jeroen van de Kamp, who originated VRC a year ago.
The comparison was designed to measure how the three hypervisors performed when running Windows XP and Vista virtual machines for Terminal Services and VDI environments.
I filled in the Microsoft employee on the news that performance almost doubled on both XenServer and vSphere, while it went up 154 percent for Hyper-V. Then I asked him if those results were in line with what he expected, and he said that they were, adding that VMware had been telling some of his customers "something completely different," and he just wanted to have his facts straight.
When I asked him if he would anonymously tell me what VMware was telling his customers, he said that they were claiming that Hyper-V does not scale as well as VMware, so users will need many more physical servers to do the same workload. If this is true, the Microsoft employee said, it is a valid point since doubling the number of servers increases acquisition costs, physical/environment costs -- meaning, space -- and power cooling costs. However, he said, VMware's claims did not appear to be true based on any objective benchmarks/data he had seen.
Closing with, "I am not sure if this is a loose cannon" or "company-wide FUD," he said he just wanted to be sure that he was being accurate.
These are the kinds of claims that are swapped back and forth by competitors all the time, so they come as no surprise. What would be interesting to know is, if any of you readers have heard the same claims. So, have you? Comment here or send me e-mail.
Posted by Bruce Hoard on 03/05/2010 at 12:48 PM