Novell

There was a time when was pervasive in the enterprise. It had far and away the best directory management and security model. What happened to Novell has, I think, parallels in the virtual space right now.

Consider: Novell was the platform of choice for enterprises, in the then-emergent area of identity management. VMware is the platform of choice now for enterprises in the area of virtualization. There are many other parallels: perceived as expensive but low-career-risk choices, jacking up prices for a user base who are vested and have little short-term choice, complex licensing models, strong lock-in, Microsoft snapping at the heels and shipping free product to take market share.

As far as I can see, VMware needs to change at least one thing in the near term in order to avoid the fate that befell Novell. has far and away the best management interface of any virtualisation system. is cool but a way behind, the stuff is OK but suffering from being moved effectively into the private space from open source thanks to Citrix,  is based on System Center, which is Microsoft's answer to, ,  and so on, with all the complexities that implies. A framework in which you can build something good, but it will require significant investment. VMware's vCenter is, right now, clean, reaonsably stable, and extensible. We can plug in Update Manager, vendor-specific storage, hardware status and so on. Everything the busy virtual operations team needs, with one massive exception: non-VMware hypervisors.

Why is that important? Three reasons (at least). First, the need to manage diversity. Top-down enforcement of standards has never really worked well, there is always a certain irreducible level of complexity. Someone somewhere will have a legacy Solaris application. Second, response to business demands. If the business wants to deploy or  then we, as service opriented IT providers, should be in the business of enabling that, not trying to push them to use our preferred products and hypervisors. Third, while VMware currently has arguably the best hypervisor, it is a highly dynamic field with many awesome brains applied to it. In the near term we will, I am sure, see hypervisors embedded in server hardware, as they are in some desktop hardware now. Those hypervisors may well not be VMware.

So in my view if VMware does not open its doors to managing other hypervisors, it will become the Novell of virtualization. Or the Betamax, if you prefer. Technically superb, but outpaced in the market by cheaper and more open competitors.

The excellent Simon Wardley opines that the IaaS part of VMware is separately viable; this is a good observation and invites speculation of a future divestiture, but the writing is, according to some, already on the wall for the hypervisor.