Watching Paul Maritz and Todd Nielson run VMware is giving me a serious case of déjà vu. In 1995 and 1996, I was Gartner’s Research Director for the Windows System Software platform (the Windows OS – Server and Desktop). Windows 95 had just taken the world by storm, and Windows NT was just getting going. I had written the research note telling Gartner’s customers that OS/2 was dead, that they should not buy any more of it and that they had three years to migrate away from it to Windows. Paul Maritz was in charge of the Windows development team and Todd Nielson had (in my opinion) the most important job on the Windows team. Todd ran the ISV program designed to get third party developer support (in the form of software companies and teams writing custom programs for enterprises) for Windows tools and the Windows OS.
Since that time the model for Microsoft’s success with the Windows OS has become clearly established. It is based upon the following principles:
- Having the most technically superior operating system is not nearly as important as having the broadest and widest support for your platform from third party developers. Microsoft has done an astonishing set of things very well here. Great tools for developers, great support for developers, great education for developers, and most importantly a welcoming and partner attitude towards anyone who added value to the Windows platform in any way. This welcoming and partner attitude was extended to everyone – even if they competed with some other aspect of Microsoft’s offerings. This continues to this day and remains the single greatest strength of the Windows platform.
- Establishing the high volume/low price/market share leadership point in the market. Low prices, broad distribution channels, aggressive marketing, and that killer library of Microsoft and third party based applications made the Windows platform into the market share leader on the desktop and the server.
- Having a platform business model. This is a corollary of #1 and #2 above, but it is the single most important aspect of the success of Windows. There was a very simple rule at Microsoft. If it was good for the success of Windows, then it got done. As Bill Gates once said, “All of our wood is behind one arrow”. Microsoft was never confused about what business it was really in. It was (and is) in the Windows platform business, and everything else is just a supporting member of the cast.
- Good enough (although not always great) forward compatibility for applications. DOS applications were supported in Windows. 16 bit Windows applications were supported in 32 bit Windows. 32 bit Windows 95 and Windows 98 applications were (for the most part) supported in Windows NT and Windows 2000 (the issues with Windows XP applications and Vista are a painful exception here).
- Aggressive incorporation of standalone products and markets into the core platform. Perhaps the single most controversial aspect of Microsoft’s strategy, this included bundling TCP/IP networking into the OS (leading to the ultimate demise of Novell), bundling IE into the platform (leading to the ultimate demise of Netscape – and a lawsuit by the anti-trust authorities in the USA and the EU), as well as many other little things to numerous to count. As controversial as the “Microsoft snowplow” became, it was critical to the success of the platform, since it let developers know that there was going to be growing set of services that they could leverage in their applications.
Now it is VMware’s turn to attempt to become the next great platform company, lead by two people who were instrumental to the outstanding success of Microsoft Windows platform. Let’s evaluate VMware’s model and compare VMware to how Microsoft became so successful with Windows:
- Breadth of third party support. Microsoft built a killer ecosystem around Windows that continues to be its greatest strength. Where does VMware stand on this front? Good, but not great. True VMware has a great VAR community that thrives off of doing VMware implementations for customers large and small. But the executives that run these companies sleep with one eye open because they just do not trust VMware completely as a partner. As far as third party software/hardware vendors are concerned the story is again mixed. True that after a small blowup in the blogosphere VMware reassured everyone that all third party vendors who want to pay for a booth at VMworld will in fact be welcome. But if we dig into VMware’s relationships with these third party vendors we find that VMware does not cooperate with them very well in field sales situations and that VMware often says “gee thanks for the idea” and puts the latest innovations from third party “partners” into its own products. Being owned by EMC does not help here either as it creates the perception of an unfair playing field in the storage arena. The bottom line here is that neither the third party VAR community nor the third party vendor community feels a great deal of loyalty to VMware. This is very unlike the passion that Microsoft was able to engender for the Windows platform as it was establishing its dominance in the late 1990’s.
- Establishing the high volume market leadership position. It is clear that VMware is currently the market leader. But it is also clear that VMware is just a vulnerable as Unix was when Unix was in that position in the server operating system market (by Unix in this case I mean the combination of HP UX, IBM AIX, Sun Solaris, SCO, etc., not today’s Linux). VMware’s recent pricing moves with vSphere have made this problem worse on two fronts. VMware has become more expensive to the very large and sophisticated enterprises who want and need all of its advanced features, and it has not become nearly inexpensive enough to capture the high volume position in the rapidly growing SMB/SME market. Executives in large enterprises are counting the days until Microsoft arrives with a competitive Hyper-V offering, and the SMB/SME market – much of which is very Windows oriented to begin with is happy to take Microsoft’s offer of virtualization as a free feature of the OS and run with it.
- Forward compatibility. Since VMware really did not have much in the way of third party API’s prior to vSphere there is little to go on here. The fact that vSphere is 64 bit only will create a problem for some enterprises who implemented ESX on 32 bit hardware. But VMware cannot be faulted for going 64 bit with vSphere and taking advantage of what Intel and AMD were providing. The interesting story here is just starting. With vSphere, VMware has new vStorage, vNetwork, and VMSafe API’s. How it treats the third party companies that could and should take advantage of these API’s, and how they are protected and partnered with on a forward going basis will determine a large part of VMware’s future success.
- Cannibalizing adjacent markets. While VMware desperately needs support from the VAR and ISV communities, it must at the same time fight an extremely aggressive war in the system software market against Microsoft, Citrix, Red Hat and Oracle. Microsoft is trying to package VMware out of existence by including virtualization as feature into the OS. VMware must return the favor by making the OS into a feature of its virtualization platform. Success at this will require that Todd Neilson will have to once again do what he did so well at Microsoft – evangelize a set of API’s (that for the most part do not exist yet) to developers of product and custom solutions that eviscerate the role of the operating system as we know it today.
- The platform based business model. This is where I think VMware is most confused. I think all of the talk about free hypervisors and value added management offerings has gotten VMware seriously off track. VMware needs to establish vSphere as the dominant virtualization platform in all segments of the system software business. This means making the platform the standard from the SMB to the largest enterprise and making the success of the platform (and not the surrounding tools) the primary goal of company. This means welcoming and partnering with anyone and everyone who helps establish the success of the platform even if it means that they compete with some other value added tool. Finally in means that to the extent that any value added tool is offered and marketed, those tools need to be open and extensible so that third parties can add their value to them.
In summary, VMware is today a very successful product. It is not yet, by any stretch of the imagination a successful platform. Paul Maritz and Todd Neilson have done this once before. Can they modernize the Windows playbook for the market conditions of today and do it again? Right now, the odds do not look so good.