DISCLAIMER: Yes, I’m a Microsoft MVP, MCT, MCITP, CCNA and hold various other certifications, however when I write articles I will take the view that there might be a better technology out there than Microsoft, but I will ensure I write articles with the proper facts being portrayed and not try to put a “media” spin on things.
OK, now that’s over, let’s look at what has got me revving. A bloke called Elias Khnaser has written an article for Information Week entitled “9 Reasons Enterprises Shouldn’t Switch to Hyper-V”. He didn’t put a disclaimer at the top, but Elias works for a company called Artemis Technology and if you go to their “Partner” page, Artemis is a VMware Enterprise Partner and they consider this their one of their “Spotlights” compared to their Microsoft Gold Partner status which is just listed as a valued partner. (UPDATE: since writing this article, the logo has changed to include their areas of expertise and has been fixed, however at the time of writing the logo was “warped” and furthermore if you moused over it, the description about the partner was set to “information to come”, whereas all of the other valued parnters seemd to have descriptions).
Anyways, let’s look at the article…
1. Breadth of OS Support….:
Hyper-V, however, supports only Windows and SuSE Linux
Microsoft support more than just Windows OS’s and SUSE. Here’s another page I’d like to steer Elias to:
OK, so it’s a piece of software, but what does it say? Yep, that says of the top 15 OS’s 12 are Microsoft’s, so why support such a breadth of OS’s if there’s no need to…Anyways, I digress. Let’s carry on:
2. Memory Management
In this article it goes in to Memory management and talks about how Microsoft just say throw more memory at the situation whereas he states VMware can overcommit and utilise more memory…Interesting, in contrast to this article:
That white paper clearly states:
Avoid frequent memory reclamation. Make sure the host has more physical memory than the total amount of memory that will be used by ESX plus the sum of the working set sizes that will be used by all the virtual machines running at any one time. (Note: ESX does, however, allow some memory overcommitment without impacting performance by using the memory management mechanisms described in “Resource Management Best Practices” on page 12 [of this document].
key word of course is *some*, yet everyone knows you never overcommit memory in a production environment (thanks to my friend Mitch Garvis for the heads up on this one – A Brief Discussion of Security with Regard to Resource Over-Commitment in VMware)
Well, don’t get me started on this one as VMware has a kernel infrastructure that means if you inject one malous driver in to the Hypervisor layer it can (and will) affect EVERY VM you have. Hyper-V does it differently. Here’s a reference for the differences:
Biggest difference is microkernalised hypervisors versus monolothic hypervisors.
4. Live Migration
Well, lets look at this one. In order to do it with VMware, it’s not as straight forward (oh wait, it’s not mentioned in this review of *one* paragraph) as it seems either.
5. Priority Restart
Seems as though the spin on this paragraph is going down the clustered route not a priority restart route. He mentions Exchange, IIS, SQL all of which, you don’t want a VM infrastructure to *move*. You want them highly available via clustering, not a VM management utility or tool…
6. Fault Tolerance
Not sure where this one is heading, but again it seems like he’s letting VMware control the applications, something any good system administrator (see third party software and reliability).
7. Hot Adds
All I need to say is CSV and I don’t mean comma separated values.
8. Third Party Vendor Support
Please list some…Furthermore, I’d ask why (and 9 is maturity) if VMware is so much better, do they need third party products to make their product good? Hyper-V has SCVMM R2 and that’s all you need, period.
Sure Hyper-V hasn’t been around long, but you have to admit, it’s gaining ground on VMware at a very fast pace now that the R2 version is out and the “kinks” have begun to be ironed out…
Morale is, it seems this story has a load of FUD in it and that proper research wasn’t done in order to make it impartial. Next time an article like this is written, maybe it should be prefaced with the caveat the author is a VMware addict or seems to be trying to have it out for Microsoft Hyper-V, for whatever reason that might be.