A recent TechTarget piece pits VMware vSphere memory management technologies against the new Microsoft Hyper-V Dynamic Memory. While certainly an interesting topic, I was disappointed by some of the inaccurate statements propping up the Hyper-V side. With no facility to provide comments on that article directly, I thought I’d take a moment to set the record straight.
The key claim — that Microsoft offers more control over virtual machine memory — is misinformed at best:
Hyper-V Dynamic Memory also has a greater range of configurable options than does VMware memory overcommit. Users can assign limits to problematic VMs with memory-hungry workloads, and if memory contention occurs, users can prioritize specific VMs. A configurable buffer value also identifies how much extra memory is reserved for short-term needs between rebalancing passes.
Readers seeking to find the true differences between these platforms will need to search elsewhere. A factual comparison reveals that Hyper-V Dynamic Memory offers no advantage over VMware vSphere: A VMware ESX VM has settings for memory size, limit, reservation, and shares to specify priority. Not only that, VMware vSphere offers a comprehensive range of memory management technologies: ballooning, page sharing, compression, and host swapping.
What vSphere doesn’t have is a reserve memory buffer setting — this is an artifact of the Dynamic Memory design, accommodating for lag time inherent to the hot-add process. Interestingly, when running important enterprise applications like SQL Server, Microsoft recommends cranking the buffer down to the lowest possible setting.
VMware vSphere offers the widest range of memory management and configuration capabilities, accommodating even the most demanding workloads.
Wow, I am shocked to see how TechTarget can draw a so flawed conclusion.
First, to borrow a phrase from Winston Churchill, some vendors engage in quite a bit of terminological inexactitude to market inferior products.
Second, what often poses as investigative reporting is really just a regurgitation of press releases and marketing material.
I think you missed the point. Techtarget made no stance on these pieces. These come from two different “Contributors” from two different reference and experience points. I agree that some of the wording in the Hyper-V article is a bit aggressive in its proclamations of superiority, but I see you fail to be equally as (V)critical of the equally vague stabs at Dynamic Memory from the VMware article like:
“Hyper-V Dynamic Memory has some big problems”
and
“Microsoft’s memory management approach is very weak; the company should have once again copied what VMware does. If you’re going to innovate, do it right. Otherwise, just bite the bullet and emulate.”
These articles are perfect opinionated contraries of each other and have done exactly what Techtarget had hoped, spur on the rivalry.
“Even worse, you can’t hot-remove RAM from a VM with Hyper-V Dynamic Memory; you must reboot a VM to reduce the amount of memory. Adding or removing memory from any running server is a bad idea”
What!?
Last time I checked, you can’t hot-remove RAM from either platform, hence the balloon driver pushes the non-required RAM out and thus makes it available elsewhere.
What is he on about?
As Rob said, SearchServerVirtualization.com did not take a stance on VMware memory overcommit or Hyper-V Dynamic Memory in this article. It was not an article on “Why Dynamic Memory is better than memory overcommit.” It was a face-off between two contributors who took different sides and shared their opinions.
Readers looking for the pro-VMware side did not need to “search elsewhere,” as Eric Gray wrote; that information was on the same page, in Eric Siebert’s portion of the face-off. The goal of this face-off, as with all of our articles, was to present all sides of a particular issue and let our readers decide what is best for them.
Colin Steele, Senior Site Editor
SearchServerVirtualization.com