Symantec Virtual Appliance Stymied by Windows Licensing

Last week Symantec declared that it would delay delivery of its security virtual appliance due to Windows CE licensing restrictions. As part of the announcement, product manager Gary Sabala indicated a desire to “move to more open source components” in order to “ease our ability to work around some of the licensing issues.” Not surprising.

Microsoft has no reason to endorse virtual appliances as an application delivery format because they eliminate the general purpose operating system's value proposition. Also not surprising that Symantec has reached the conclusion that they need more control over the components of a solution that ships as a virtual appliance. Having your product shipments subject to the licensing whims of an erstwhile “partner” is not a good idea.

I could drone on and on (as I typically do) regarding the deficiencies of the general purpose operating system as a platform for virtual appliances, but today I will point you to some of the hits from the past that hammer home this point:

Your Application is Your Avatar – why settle for the limitations imposed on your application by the requirement to connect to a physical server?

Microsoft Attempts Software Appliances – the key word is “attempts,” as the Microsoft packaging approach for the operating system is woefully inadequate due to its general purpose operating system legacy.

Yrtsudni Erawtfos – Putting the operating system first among customer application considerations is the business equivalent of spelling “software industry” backwards.

Microsoft Seeks Pepto-Bismol Patent – Although not yet delivering applications as virtual appliances, they are nonetheless seeking intellectual property that paves the path to the future.

Now it's time to place those sales calls to Intel and Symantec. I think they might be in the market for a real virtual appliance platform. . .

VMware IPO Sizzle

The financial markets have voted with their wallet, and the VMware phenomenon is for real. With a market capitalization at the close of trading yesterday of about $17B, the VMware franchise is more valuable than Sun and about 5 times more valuable than Red Hat. If you thought Linux was a hot trend, the hypervisor trend is going to be even better (note that XenSource was acquired by Citrix for $500M in the wake of the VMware IPO). But what makes VMware so valuable?

VMware offers the possibility for the technology industry to escape the miserable complexity of the general purpose operating system. Microsoft and Red Hat have established valuable franchises by offering a “standard” as the solution to general purpose operating system misery. The theory of the “standard” is that complexity is reduced by eliminating variability. The problem with this “standard” theory is the painful reality that each software application and hardware combination actually requires a different implementation of the “standard.” It is impossible to run multiple applications on a single copy of a general purpose operating system because each application requires different configurations of the “standard” in order to work properly. It is also impossible to quickly move applications to a new hardware platform because the “standard” that ran the application on the old server needs to be “upgraded” to run on the new server because the drivers are different. That “upgrade” inevitably “breaks” the application, resulting in an expensive re-stabilization process just to move the application to a new server.

In the new world where VMware is the king of the hill (and therefore collects the kings ransom for owning the hill), applications are completely de-coupled from each other and from the hardware. Each application is a virtual appliance with Just Enough OS (JeOS or “juice”) to support the application and attach to the hypervisor. Want to run multiple applications on the same machine? No problem. Each virtual appliance runs autonomously without affecting other virtual appliances on the machine. Want to upgrade the server? No problem. Fire up the new box and “vmotion” the old application over to it. Less than a millisecond for the migration. No porting. No complexity. No more general purpose operating system.

VMware is worth $17B because their technology enables the dream of utility (or grid) computing. By eliminating the operating system as a bottleneck for attaching applications to computers, applications can now run flawlessly on any computer that has the VMware hypervisor as the bottom layer -- with no expert resources required for testing, porting, stabilization, etc. How big can VMware get? What is the value of owning the operating system market PLUS the value of removing wasted administration effort associated with stabilizing applications to run on the “standard” general purpose OS? Big. Very big.