Tuesday, August 18, 2009

Comparing Hypervisors

Last year I posted a blog entry comparing VMware ESX Server and Microsoft Hyper-V. Since not many companies had been able to perform a head to head comparison and gone through (or attempted to go through) the process to do it, I had gathered some performance data from a few different tests that were run on similar hardware and was able to draw some comparisons between the two Hypervisors.

Looking back, I feel that I need to update this blog, but instead of updating the original blog, I am posting this new entry.

Since this blog post, there have been a few comparisons that have been posted that do have head to head comparisons with VMware ESX Server and Hyper-V. Project VRC (Virtual Reality Check) has made some comparisons based on Desktop Workloads, and Virtualization Review posted an article comparing the different Hypervisors as well. Now some of these are contentious, and the benchmarking guidance that VMware gives is important as seen by their response, but the fact of the matter remains, they Hypervisors performed comparably.

I was looking at my previous blog, and reading the comments, and I think this blog should be updated. Some people, in the comments, are saying I am comparing Apples with Oranges in this post as there were a LOT of variables that were different between the two runs. The main differences (as noted at the end of the blog, and in the papers) were that the computers were very similar (same processor and same amount of RAM), but different (different numbers of DIMMs, and possibly different bus speeds), AND that the versions of vConsolidate were different. With the aforementioned, and the release of vSphere 4 and Windows Server 2008 R2 Hyper-V, the comparison below is probably not correct today. If you compared the versions mentioned in the post, you would probably see similar results, but BOTH companies have made major improvements in their Virtualization Technology, so take this blog for what it was (a sampling at the time), and hopefully, we will see some performance data comparing Microsoft and VMware in an independent Head to Head comparison. When we do, what we will probably see is that both products will beat the other in some capabilities, and lose to the other in others.

What really matters is that ANY company can benefit from virtualization NO MATTER WHICH vendor they choose, and if you aren't virtualizing START.

No comments: