by
Kevin Goodman
I had the pleasure of attending BriForum last week in Chicago. BriForum is always a great time with knowledgeable attendees (and exhibitors). One topic of interest was the revelation that Paul Maritz is leaving VMware at the end of August to be replaced by Pat Gelsinger.
There is plenty of speculation as to why the move took place. Some news outlets are suggesting that this clears the way for Maritz to take over as CEO of EMC when Joe Tucci retires. Others are treating it as a termination, with Maritz’s transfer to EMC is a polite way of letting him “ride off into the sunset.” I have no inside information (although I do hold a position in VMW stock) as to the real reason, but I can imagine either case being reasonable. Past executives at VMware have “gone on sabbatical,” “off to do special projects,” or the equally vague “taken a step back”—all never to be seen again. Of course if Joe Tucci really is retiring all of this hand waving might be necessary since the news of Maritz’s departure was leaked. Indeed, Wikipedia lists Gelsinger as taking over as CEO effective September 1st.
In any event, Mr. Gelsinger now has an opportunity (or perhaps the mandate) to make sweeping changes at VMware. I know of no CEO who replaced a non-retiring CEO as having a mandate to *not* change. I doubt you’ll ever hear a Board of Directors say, “We fired the last guy, but don’t feel pressured to do anything different.”
Many BriForum attendees asked me what I thought was going on since I worked in the End User Computing group at VMware from 2010 until earlier this year. I’ll give you my thoughts on what may happen, but first I'll give some context so you can understand terminology and such.
Understanding VMware's product business
VMware’s products fall into three categories. (VMware likes to call them layers in a cake or the “New IT Stack.”) The top layer is end-user computing, the middle layer is the cloud application platform, and the bottom layer is what VMware calls “the underlying cloud infrastructure and management." (e.g. server virtualization, management, and monitoring.)
End User Computing products include Zimbra, SocialCast, Horizon, View, and a few others. While I was at VMware I spent my time figuring out ways to make VMware View better. I just assumed given VMware’s stock price that the other two “layers” in the cake were doing fine. However, while researching my “What Two Years in the Trenches have Taught Kevin Goodman about VDI” BriForum presentation for London last May, I decided to pull VMware’s jacket and see just how much money View makes. The following table is from VMware’s 2011 Annual Report:
|
|
VMware, Inc.
|
|
|
|
NOTES TO CONSOLIDATED FINANCIAL STATEMENTS (continued)
|
|
|
|
|
|
|
|
(thousands):
|
|
|
|
|
|
|
|
|
|
|
|
|
For the Year Ended Dec 31,
|
|
|
|
|
|
|
|
|
|
2011
|
2010
|
2009
|
|
|
Cloud Infrastructure and Mgmt
|
$1,665,599
|
$1,263,232
|
$945,018
|
|
|
|
|
|
|
|
|
Other products
|
175,570
|
138,192
|
84,424
|
|
|
|
|
|
|
|
|
License revenues
|
1,841,169
|
1,401,424
|
1,029,442
|
|
|
|
|
|
|
|
|
Services revenues
|
1,925,927
|
1,455,919
|
994,495
|
|
|
|
|
|
|
|
|
Total
|
$3,767,096
|
$2,857,343
|
$2,023,937
|
|
Apparently, VMware does not separate out the revenue for View. As you can see from the table VMware lumps the top and middle layers into the catchall, “Other Products.” Together the Cloud Application Platform and End User Computing managed to sell $175 million in 2011, or less than TEN PERCENT of the company's total revenue.
If I'm the new CEO, this is probably the second thing that catches my eye. One leg of his three-legged stool produces 90 percent of the revenue. Maybe that's not a big deal if the other two legs are growing at a rapid pace, but “Other Products” only grew 28% year over year from 2010—that’s lower than the 31% the company grew overall.
I said the three-legged stool was the “second” thing I would want to know because the first thing would be of the $3.5 billion dollars VMware earned in 2011 for Cloud Infrastructure and Management (including services), how much is in “Private Clouds” versus “Public Clouds?” Private Clouds are what you and I would call Data Centers. Public Clouds are what you and I would call products like Amazon Web Services’ EC2 platform and Microsoft Azure. VMware pretty much owns the Private Cloud Space According to VMware’s customer list. However, from that list it appears that everyone who needs data center virtualization already has it.
The word on the street is that EC2 is kicking VMware’s butt in the Public Cloud space, and in case you didn’t know, EC2 is built on top of Xen—the open source hypervisor. (Not to be confused with XenServer the product from Citrix). The ‘E’ in EC2 stands for Elastic, and this is a wonderful feature for companies like NetFlix that need more servers on Sunday night to manage the load when we are streaming movies than they do on Thursday morning when you and I are at work. Amazon Web Services only charge for what you use and need. Compare this to a data center where once the server consolidation takes place, it typically only grows or stays static. (Now before you disparage anyone from VMware for not recognizing this, remember that ESX started in the data centers. Amazon created EC2 from the base XEN and was free to innovate without worrying about supporting existing data center customers.)
All public company CEOs know that stock price is based upon revenue growth, and I’m guessing that the growth will be found by tapping into the public cloud market. If so, look for Gelsinger to give the “We’re-Going-Back-To-Our-Roots-Speech” and announce a concentration on improving the Cloud Infrastructure Layer. If that is the case, a decision has to be made on the other two layers: Continue investing, or cut your losses and run?
Some web sites are predicting a spin-off of the middle layer, which has products like SpringSource and Cloud Foundry, but I believe End User Computing (EUC) is an even bigger issue and needs to be addressed—and quickly to avoid any further brain drain. I’m guessing that EUC accounts for the lion’s share of the $175M revenue and therefore has the most to lose should the talent walk out the door. EUC has lots of smart people who are very perceptive and will leave for better opportunities if they believe they work in a lame-duck division. As I documented in the Foreward to “The VDI Delusion,” there is a revolving door of executives and architects at End User Computing and some of the smartest EUC minds are now at FaceBook, AutoDesk, MobileIron, Cisco and Authentic8. (To name a few. Or they've already jumped ship to work in the Infrastructure layer at VMware.) Retaining the remainder of the talent will be necessary to have a company worth spinning off.
EUC as an independent company would allow VMware to retrench and concentrate on public clouds. I don’t believe that this it too far of a stretch to imagine. Currently, when View sells a virtualized desktop, it means a copy of ESX in the data center. When the main competition, Citrix, sells a virtualized desktop, it means—wait for it—another copy of ESX in the data center. That’s right. The majority of XenDesktop deals end up on VMware servers. So VMware shouldn’t care where their ESX revenue comes from.
As an independent company, EUC generating $175 million is a nice small sized company. It would be free to innovate without the burden of having to become a billion dollar company in two years just to meet VMware’s revenue goals. They could immediately start supporting RDS without worrying about it cutting into infrastructure sales.
Will any of this happen?
Who knows, but one indication could be the dreaded B-word: “Bundle.” A bundle of all products into one super product could be an indication that VMware is considering a spin-off or outright sale of EUC to another company. It would go something like this: VMware makes an announcement that for clarity's sake from now on, all End User Computing products are branded as View.
Instead of Zimbra, it would become View Mail. Instead of Horizon you would have View SaaS Manager. SocialCast could be renamed SocialView and so on.
It should be an interesting VMWorld this year.
(Note: You must be logged in to post a comment.)
If you log in and nothing happens, delete your cookies from BrianMadden.com and try again. Sorry about that, but we had to make a one-time change to the cookie path when we migrated web servers.