I always like it when technologists have a fight. Not a proper fight you understand, one with bare fists and the like, but a good old-fashioning mauling with words. Which brings us to the blog titled “A Stronger Case For Virtualizing Exchange Server 2013 - Think "Performance” published on VMware’s site by a person named Deji Akomolafe, who turns out to be a "CTO Ambassador at VMware", according to their LinkedIn profile. As such, I assume that their views are endorsed in some way by VMware.
In any case, the post begins with:
“We have been aware for several years that Microsoft's sizing recommendation for Exchange Server 2013 is the number one cause of every performance issue that have been reported to VMware since the release of Exchange Server 2013, and it is gratifying that Microsoft is acknowledging this as well.” This statement amused me because my experience is that many problems with virtualized Exchange deployments arise through the relative complexity of the environment or poor execution of plans. However, I guess VMware knows the data about the problems that are reported to them.
The rip-roaring read then goes on to inform us that an EHLO post last month acknowledges that CPU over-sizing is “one of the chief causes of performance issues on Exchange Server 2013” and that “the Exchange 2013 Server Role Requirements Calculator is the main culprit in this state of affair (sic).”
As I understand the argument, VMware is upset that the calculator recommends customers to deploy too many CPU resources and that “unfortunately, not may [many] customers like to buck Microsoft especially in the face of strident claims of "This is our product and we are the experts." There's some truth here because there are many true believers working for Microsoft in the field who would never disbelieve anything that emits from Redmond. But professionals know when to take advice and when hot air is blowing...
I’m sure that Ross Smith IV, who is responsible for the Exchange 2013 server calculator, is crying into his cornflakes at the indignation flowing from VMware. I mean, seriously, how many people who deploy Exchange 2013 in any sort of large-scale manner (which is where a tool like the server calculator is most useful) will take the output from an Excel worksheet and say “Eureka – now I know what hardware to order!”. Well, maybe there are a couple... but not too many.
Most experienced people take the output from any general-purpose sizing tool and cast a cold eye over its recommendations to put them into context with the operational and business requirements for a deployment. In other words, the recommendations are adjusted. And yes, sometimes those recommendations are adjusted to make sure that Exchange 2013 works well when deployed on virtualized servers, either Hyper-V or VMware.
The blog then goes on to say “Microsoft has evolved in some ways and the Exchange team is now more open in discussing the inherent defects in Exchange Server 2013”. Quite. I am sure that Satya Nadella is delighted to know that Microsoft has evolved (but only in "some ways"). Even more so, it’s good to know that the Exchange team has turned the corner and is able to discuss the inherent defects that exist in Exchange 2013. After all, Exchange 2013 has been available for almost three years so it’s about time that those wicked engineers should acknowledge the rubbish they have been peddling to the market.
Jeff Mealiffe, who is normally well-respected when he speaks about performance matters for Exchange, is skewered because he continues “to push the "Combined Role" design recommendation, in spite of the fact that such design unnecessarily complicates performance troubleshooting and hinders fault domain isolation.” Combined role means a multi-role server, which I think that every reasonable expert working with Exchange has concluded is the only way to go because it increases server utilization and improves the overall resilience of any deployment. But it’s a bad thing in VMware’s world, which is a pity for them because Exchange 2016 only supports multi-role servers, so I guess they will just have to get used to that fact.
The blog says that Exchange team is “not virtualization friendly”, an allegation that must come as news to the many companies who happily run Exchange 2013 on VMware or Hyper-V. Maybe it’s the fact that Microsoft has been driving Exchange to run on commodity hardware and low-cost storage that is causing VMware most heartache when customers discover that on-premises servers run just fine on low-spec servers, even if this is evidence of “Microsoft's insistence on multitudinous proliferation of "itsy-bitsy-sized" silo'ed physical hardware for Exchange Server”. Perhaps VMware would prefer if we were still back in the world of Exchange 2003 or Exchange 2007 when SAN-class storage was the norm – storage that can accommodate scads of virtual servers, no?
I could go on, but really… A couple of points come to mind.
First, Exchange does just fine in virtualized environments providing you do it right. That means that you have sufficient experience and expertise to take whatever recommendations calculators or any other sizing tools spit out and adjust them for your circumstances. It’s just fine if you want to increase the CPU load on all your virtual Exchange servers providing that you understand the trade-offs in terms of resilience and high availability. Go for it and have a ball. Me, I find it interesting that not one virtual server is used inside Exchange Online. It’s all low-cost commodity hardware and cheap JBOD storage. Just saying…
Second, blogs like this make me wonder whether companies like VMware have started to realize that the pool of available customers that might want to virtualize Exchange is shrinking due to the growing share taken by Office 365. If VMware can show that virtualization is cheaper than Microsoft’s sizing tools might make it out to me, customers might be more willing to consider going virtual instead of going into the cloud. If you do go virtual, make sure you go in with your eyes open. My experience is that quite a few companies have experienced problems with virtual environments because of the extra complexity required in deployment, operations, and management. But done right, virtualization is certainly an option to consider.
Third, wouldn’t it be better use of VMware’s time to publish well-argued and pertinent observations on how you can take the output of Microsoft sizing tools and adjust them for their platform rather than bellyaching about topics they cannot control. It is perfectly OK for Microsoft to recommend a preferred architecture, that multi-role servers are preferable, and that certain sizing guidelines should be followed. They are, after all, the responsible party for Exchange. VMware can’t change those factors. They can change the perceptions of the Exchange community by arguing a case in a more coherent and cogent manner than found in the “Virtualize Business Critical Applications” blog.
But then again, it’s great when real feelings about technology surface and burst out into public. So thanks VMware… now go back and publish some good stuff please. Like how to take Microsoft's Preferred Architecture for Exchange 2013 (and 2016) and make servers fly on VMware.
Follow Tony @12Knocksinna