Microsoft already took 23% of virtualization market says IDC - UPDATED

Posted by Alessandro Perilli   |   Thursday, October 16, 2008   |   19 Comments

idc logo

Today the analysis firm IDC released some surprising numbers about virtualization vendors market share:

  • VMware: 44% (combining ESX and Server)
  • Microsoft: 23% (combining Hyper-V and Virtual Server)

Isn’t clear how much of these market shares translates in production deployments and how much is about test & development or shelfware, but one thing is for sure: the Microsoft gain is remarkable (and its growing case studies library confirms it) even if VMware still leads with 78% revenue share.


Besides the numbers above there is something else that is interesting: worldwide virtualization license shipments in the second quarter of 2008 (2Q08) increased 53% year over year, compared to a 72% year-over-year increase the previous quarter.

Who ships more servers for virtualization duties? HP (34% market share) followed by Dell (25%) and IBM (16%).


Update: Easy to guess the report above generated a huge number of reactions (and complains from VMware).

Specifically IDC is questioned on several obscure points:

  • In its previous report about virtualization market shares, about Q1 2008, IDC assigned 20% to Microsoft.
    Hyper-V was formally released on June 26, 2008, which is two days before the end of Q2. So it seems unlikely that Hyper-V could get 3% in just two business days.
    As some virtualization.info pointed in the comments it’s very likely that IDC included in this count also beta and RC deployments. If so the numbers couldn’t reflect the real state of the market because VMware tents to keep ESX betas private (like for the upcoming ESX 4.0).
  • Unlike the past, this time IDC didn’t compare its findings with VMware and refused to break out the reported numbers for each product.
  • There’s an incongruence between the VMware estimated year over year growth (52%) and the real one reported by VMware to the SEC (39%) for the Q2 2008.

For all these reasons we expect an official statement from IDC clarifying the methodology that provided the market shares.

Meanwhile it’s worthwhile to answer a question that Mike DiPetrillo raised in the comments: why a 23% market share for Microsoft should be considered remarkable? The answer we can give:

  • Because it didn’t go lower despite ESX 3.5 Update 2 introduced a huge advantage in terms of features and most VMware efforts in marketing is focused on highlighting such difference
  • Because one would assume (maybe incorrectly) that customers started to dismiss Virtual Server in the last few months. In such case Hyper-V may have some more than 3% market share.

Labels: , , , , ,

19 Comments

Anonymous Mike DiPetrillo Thursday, October 16, 2008 10:06:00 PM  
Alessandro,

You might want to go back to the previous IDC report and take a look there. Microsoft already had 20% market with just Virtual Server. The new combined Virtual Server and Hyper-V number is only up 3% to 23%. It's not that remarkable at all really.
Anonymous Alberto Farronato Friday, October 17, 2008 12:44:00 AM  
Alessandro,

IDC market share estimates refer to Q2 2008 which ended on June 30th. Hyper-V RTM was announced on June 26th! It is unrealistic that such a strong uptick happened in just 2 business days, even if Hyper-V was a tremendous success. It seems to me that IDC's survey is including Hyper-V Beta testers, which is a flawed methodology to determine market share.
Anonymous Anonymous Friday, October 17, 2008 6:08:00 AM  
Hyper-V RC1 came out in May. Lots of people were running it by then.
Anonymous Massimo Re Ferre' Friday, October 17, 2008 6:23:00 PM  
If Mike got it right in the first post (I didn't check but I trust Mike) the previous report gave already MS a 20% share. The current 23% marketshare does include the Virtual Server figures so it makes it very difficult to depict Hyper-V success... it could be that they all (23%) migrated to Hyper-V in 2 business days (or since RC-1) or that Hyper-V still have 0% of marketshare and Virtual Server has been growing 3%....

We just can't say.

Massimo.
Blogger Migue Friday, October 17, 2008 9:41:00 PM  
I love VMware and I love Microsoft but i feel VMware is the new Netscape. The war is money and MS knows very well how to win.
Anonymous Mike DiPetrillo Friday, October 17, 2008 10:02:00 PM  
"Hyper-V RC1 came out in May. Lots of people were running it by then."

So we're counting beta shipments as real product shipments now? Interesting math. Skews the results a little since VMware has mostly private betas and Microsoft has very public betas.
Anonymous Anonymous Friday, October 17, 2008 11:56:00 PM  
I just don't get it. I'm heavy into the virtualization market, I interact with the majority of players in the market, and I like to think I have a pretty good grasp on the lay of the land. Where in the hell have they been hiding 23%? I've seen maybe 2 instances in 2 years! Maybe it's like notepad? "90% of all computers users use notepad!". No, it's on 90 percent of computers, making available is not the same as using. There's some slight of hand here. Maybe they're bloating numbers with 2008 installs? No one takes hyper-v seriously that I'm aware of, certainly not anywhere that requires stability....I need some of whatever they're smoking. I just don't get it.
Anonymous Attila Bognár Saturday, October 18, 2008 11:45:00 AM  
Many are comparing VMware to Netscape.
Yes, it can be done, why not, but we should remember that there is also Oracle/DB2/... vs MS SQL Server...
Anonymous Anonymous Sunday, October 19, 2008 1:55:00 AM  
If MS got 20% with Virtual Server, then I would say that VMW is in a lot of trouble with HyperV
Anonymous Anonymous Monday, October 20, 2008 12:01:00 AM  
someone is playing with the numbers ! I have seen 0 installs of Microsoft Virtualisation out there !
Blogger Bjørn Anders Monday, October 20, 2008 12:45:00 AM  
Well, MS might have a 23% market share,
but they have 0% of the revenues...
Anonymous Anonymous Monday, October 20, 2008 9:10:00 AM  
actually, I'm sad to say but a major power company (ahhem, Duke Power) is in the process of transitioning *away* from VMware ESX *TO* MS Hyper-V. Someone there must be on M$'s payroll, the geeks sure don't like the idea.
Anonymous Anonymous Monday, October 20, 2008 10:05:00 PM  
As someone who has used VMware since 2.0 came out and has had VS in production and VMware server in production and now has ESX, the thing that interested me most was performance.

It was snappy.

If we were making a decision today, I would push hard for Hyper-V.

Under the workloads we use virtualization for, Hyper-V would be fine. Network World and SQL mag tested Hyper-V and found performance differences with ESX to be minimal.

Yes, we bought HA. Used once when a tech accidently turned off a server. But we never had a problem with VMware server of VS 2005 either. We switched from VS 2005 because it was a dog.

Hyper-V isn't a dog performance-wise.
Anonymous Mike DiPetrillo Monday, October 20, 2008 10:28:00 PM  
Migue, If you're going to do the whole Netscape comparison (which is a bad one since the tools on top of the underlying virt layer are not commoditized) then you do have to realize that the main person fighting that Netscape battle at Microsoft now heads up VMware. I think that gives VMware the inside track on that whole strategy. :)
Anonymous Mike DiPetrillo Tuesday, October 21, 2008 12:48:00 AM  
Here's the main points you need to know about this report from IDC:

1) This was a small sampling using just the OEM's numbers.
2) How do you count unit shipments for free products not delivered by the OEMs?
3) Microsoft only went up 5% from the previous measure.
4) VMware also went up 2% from the previous measure.
5) Hyper-V was only shipping for 2 business days in Q2 2008. There was no Hyper-V impact to these marketshare numbers.
6) VMware still commands a huge lead in actual product use.
7) There are real factual errors throughout the report.
8) This was a sponsored report with no input on revenue, shipment details, or feedback from VMware.

For more details on all of these points visit my blog at: http://www.mikedipetrillo.com/mikedvirtualization/2008/10/idc-microsoft-g.html.
Anonymous Anonymous Tuesday, October 21, 2008 1:44:00 AM  
It's also reasonable to say that whatever is reported as Microsoft's share is irrelevant as they are a proper competitor to VMware and they will provide an alternative garnering a traditional market share given how close this battle is to the OS. Let's not argue about their share but instead make sure both vendors understand what the user community needs to leverage this technology properly to save money, jobs and gain efficiency so IT ceases to be such a big burden on the typical business professional.
Anonymous Mike DiPetrillo Tuesday, October 21, 2008 2:25:00 AM  
"actually, I'm sad to say but a major power company (ahhem, Duke Power) is in the process of transitioning *away* from VMware ESX *TO* MS Hyper-V. Someone there must be on M$'s payroll, the geeks sure don't like the idea."

Not much of a case study - they've always been on MSVS. Not a switch except for old MS to new MS.
Anonymous Anonymous Tuesday, October 21, 2008 3:09:00 AM  
Mike, your blog suggests IDC is underrepresenting software revenue. Are you counting VMware Workstation in that revenue?

VMware Workstation shouldn't be counted in a discussion of server virtualization revenue.
Anonymous Ken Cline Tuesday, October 28, 2008 8:22:00 PM  
Well...I think the question is really "How many virtual machines are deployed on each platform?" - VMware is - I believe - a strong leader in the enterprise space where the features matter. In the SMB space, where there is a lack of technical skills, customer comfort with the Microsoft name will drive MANY people to use Hyper-V (just like they used MSVS, even though it was a dog).

The SMB space is also where a lot of users will download something and leave it sitting on the shelf. This is because they don't have the technical skills to implement or support it - and the don't have the time, nor desire to invest the time, to develop those skills.

As the old saying goes..."there are liars, damned liars, and statisticians" - survey results can easily be interpreted to the benefit of the survey producer...

Add New Comment