The answer is “yes” and “no”, as we shall see. And this question begs another, perhaps more important, question: How will VMware develop its business model going forward? We’ll make a guess at this too.

Why VMware is like Netscape

There is a parallel between VMware and Netscape, which goes beyond the fact that both companies appeared in the stock market with stellar IPOs. Netscape’s initial product was the Netscape browser and its stellar market valuation derived almost entirely from its initial domination of the browser market. Microsoft decided, correctly, that Netscape was undermining its grip on the PC market, so to counter the threat, it developed Internet Explorer (IE) and bundled it with Windows.

It was monopoly behavior, that was officially recognized as such by the DOJ, and it completely destroyed the browser revenue stream that Netscape was planning to develop. However, Netscape’s response to Microsoft’s challenge was exactly wrong. It tried to fight Microsoft on the desktop – a strategy that equates to invading Russia with a small band of volunteers just as winter approaches. It could have been different. Netscape could have realized that it owned the gateway to the cloud (the Internet) and challenged Microsoft with a move towards cloud computing. This is exactly what Google has done and Microsoft is now desperately trying to buy Yahoo! for billions, in a vain attempt to meet the threat.

VMware’s position is slightly different. Its hypervisor occupies a similar position to the Netscape browser, in that it has defined a “new” and important market. Microsoft has responded by providing its own hypervisor for Windows, which seems similar to its Netscape response. And when this was announced, it sent VMware’s shares tumbling, so you could think of that as similar to the launch of IE. But there is a big difference. While Microsoft really was playing monopoly when it released IE (a browser is not a necessary part of the OS), the inclusion of a hypervisor is a legitimate, even necessary, addition to an OS.

If you don’t know what a hypervisor is click on the link for a detailed explanation. Basically, it allocates out the resources of a computer, which is something that an operating system does to some degree. Microsoft is not playing monopoly by adding a hypervisor to Windows, it is extending the operating system in a logical and intelligent way. And it is doing exactly what Linux has done by adding the KVM hypervisor to the Linux kernel. It is doing what can be done using the Open Source Xen hypervisor, which both Sun Microsystems and Oracle are using. IBM has written its own hypervisor for AIX, its version of Unix, and by the way, it’s been using a hypervisor on the mainframe since Pontius was a Pilate.

The gutting of the hypervisor market

Clearly there will soon to be no revenue in hypervisors, just as for Netscape, there was no revenue in browsers. I’ve covered some of the implications of this in VMware is absolutely not a dead duck , but there’s a deal more to be said.

The first point to note is that the value that VMware delivered initially came from the fact that there were millions of under-utilized servers out there in Userland. Some of these servers were running at cpu utilization levels below four percent. Four percent is not so much an efficiency level as an inefficiency level.

The vast majority of companies addressing this problem (and most are) are now running server consolidation initiatives using VMware’s hypervisor. VMware wasn’t chosen because it was better than the competition, it was chosen because it was the only game in town. VMware had the field to itself for long enough to build some barriers to entry. So it built a complementary set of management software for its hypervisor.

It isn’t just the VMware hypervisor that is being used for the server consolidation projects. There are also up to ten VMware infrastructure products and seven distinct VMware management and automation products in use. If I were involved in a consolidation project of this kind I wouldn’t switch away from VMware’s technology just because Johnny-come-lately is offering a free hypervisor.

Microsoft, Citrix, Oracle and any other pretenders to the throne are going to have to have competitive infrastructure and management capability in order to make a dent in VMware’s market. They will eventually develop such software, for sure, but not before next Sunday.

In a year or two, VMware will not be the only game in town and most of the “low hanging fruit” of server consolidation will not be there to pick. At that time VMware will need to have developed software, which embraces all hypervisors and markets, or its server revenues will begin to dry up.

The Client Virtualization market

I’ve described the client Virtualization market in another posting (Does Client Virtualization Make Sense?). The main point to understand about it is that it’s not really a client virtualization market at all. It’s a desktop management market.

The problem that companies are trying to solve is that managing PCs is damned expensive. Cost estimates vary, but as a rough rule of thumb, the annual cost of the out-of-the-box-and-onto-the-desktop approach is around 3 times the cost of the PC itself and most of that is support cost and management cost. These costs can be cut back significantly by any of the following approaches:

• Using session based computing, as provided predominantly by Citrix, but also one or two others. This is currently the dominant approach to getting-the-PC-off-the-desk.
• Streaming a PC image to the desktop. Both Neoware (acquired recently by HP) and Ardence (acquired recently by Citrix) can do this.
• Providing a virtualized PC on a server. VMware can do this using its hypervisor and VDI (its Virtual Desktop Infrastructure) and so can one or two other vendors.
• Providing PC Blades. This is where vendors, including IBM and HP provide purpose-designed blades which are PCs and which form a pool of resources to be shared between users.

All of these approaches are competitive to some degree, all cut management and support costs and all have drawbacks of one kind or another. VMware has bought several companies, including Thinstall and Propero so that it can compete in this market.

This is a completely different market to the server virtualization market and there is no way to know how this market will evolve. It’s kinda new, even though the Citrix session-based computing part of the market is kinda old. I personally expect that virtual machine management of servers will eventually combine with the management of virtual PCs, for the sake of global optimization, to VMware’s advantage, but that won’t happen before next Sunday either.

In the meantime, Microsoft and Citrix intend to spoil VMware’s PC virtualization party and HP (and IBM too perhaps, even though it sold off its PC business) will try to establish themselves as the “desktop savior.” They have the right credentials. All of which means that VMware will meet intense competition on the desktop.

Back to Netscape

The VMware/Netscape analogy is accurate in the sense that VMware will not be able to derive long-term revenues from its hypervisor, just as Netscape could not gather long-term revenues from the browser. But the parallel runs out at that point. VMware’s stock price is no longer as stellar as it was, and it will probably fall back some more if VMware’s growth rate falls back (as I expect it to). However, VMware is not doing anything stupid like attacking Microsoft head-on. It is developing the virtualization market that it established as fast as it can, in an effort to stay ahead of the competition.

There is a requirement for an operating system that runs a whole data center. I wrote about this in a recent posting on Cisco. Cisco and/or its competitors are potential providers of such an OS and, from a different perspective, so is VMware. If it remains the lead player in the virtualization market, then it could still have a stellar future providing such software.