VDI VMware vs Hyper-V: VMware’s TCO calculator is flawed, Hyper-V cheaper and Citrix winner?
By Tarry Singh at 12 August, 2009, 11:06 am
Well this is something so cleverly dissected that I had to mention it in my post. This is an older post from DABCC but I kind of needed it as I am collecting some data to make a big case against. Clearly this takes the cake. All I notice here is that Citrix is definitely helping Microsoft build its business case where necessary and taking the whole desktop virtualization deal where they can.
I am under several NDAs and cannot talk about some 1 million VDI/Desktop virtualization deals (signed and secured) that have already been taken and we all might be in for a surprise how differently the deals have been secured with different combinations.
Do such calculations/calculators help?
Yes but beware and keep using our own brain when purchasing as you don’t want to be taken for a ride by some fancy flash calculator when making a serious competitive bid for your customer.
What is Memory Overcommit?
For a definition of overcommit, ask any employee of CitiBank, AIG, Lehman Bros etc. All banks overcommit, they take savings and lend out to others. So, we pool our memory, and those who need it draw from the pool. If everyone had to withdraw at the same time this would be analogous to a “run on a bank”, in the virtual world the COS (console operating system) starts to page to disk, and you are best off rebooting the server – because it simply grinds to a halt. Memory is expensive, but it isn’t money, Memory Overcommit seems like a good idea, until you analyse the risks. Banks are currently counting the cost of over extending themselves, I would caution System Architects to always have enough resources for any eventuality.
Other Technologies
In their notes VMware takes pains to explain that it is better than Hyper-V in various ways, however in independent testing VMware by no means comes out 33% ahead of the competition. Recent testing puts it ahead of Hyper-V, but behind Citrix and Novell Xen. There may be an argument that VMware is slighly better than Hyper-V in terms of performance, but most certainly not anything like 33%. If we eliminate memory overcommit by simply adding RAM, then the calculations are fundamentally flawed.
No related posts.
No comments yet.