Server virtualization and the products to make it happen get most of the attention in the computer business, for reasons that are currently valid but will erode as hypervisor functionality is gradually built into operating systems, system firmware or even BIOS.
Server virtualization and virtual application support will also eventually merge into the same suite of products, if only to reduce the number of hardware systems that have to be managed and the number of software layers that have to be added to make it happen.
Citrix, which has been the standard bearer for application virtualization for a decade, is pitching in to the general furor over leadership in the virtualization market with Project Kensho, a set of its own applications designed to make virtual applications run effectively on any hypervisor or even none at all.
Project Kensho acts like middleware, sitting between the applications and microkernel-based hypervisors including XenServer, Microsoft’s Hyper-V or VMware’s ESX.
The goal for Citrix, other than building on its own expertise, is to change the focus of debate in the virtualization market from whose hypervisor is best, to who can make virtualized applications run most effectively —which is more important from the customer’s point of view, anyway.
In addition to running applications, Project Kensho will, according to Citrix, allow Microsoft’s System Center Virtual Machine Manager to mange XenServer-based virtual machines as well as those riding on Hyper-V.
It’s a good idea for both customers—who Citrix execs correctly identify as being leery of being locked into just one vendor relationship, especially during an extended fight for market dominance—and for Citrix.
By taking the emphasis off the underlying server and putting it back on the applications, Citrix is trying to change the conversation about the choices IT managers face when spec’ing out their virtualization projects.
Rather than focus on whose hypervisor is faster, or whether either Microsoft or VMware will be able to offer something the other can’t, Project Kensho allows them to focus on how to virtualize applications most effectively and deliver the level of service their internal customers need.
That, theoretically, will help them get their jobs done more effectively.
On the other hand, Citrix is using the Open Virtual Machine Format (OVF)—an open standard written by XenSource and handed over to the DMTF for management, before Citrix bought XenSource.
That means the same approach Citrix is using to focus on application delivery is also open to Novell, HP, Sun, IBM or any other non-leading hypervisor company looking for an edge in the competition.
And, because the management-data exchange protocol Project Kendro uses to get Microsoft’s Virtual Machine Manager to hanle Project Kensho implementations is the open-standard Common Information Model, any other systems manager that supports CIM could do the same thing.
No one else has weighed in so far to support Project Kensho, and probably won’t, at least publicly. But any company in the data-center business that is not either VMware or Microsoft stands to benefit from the change in focus Citrix is trying to promote.
So, whether it’s in the way they package and offer professional services, bundles of virtualization-supporting applications, or in the way they advise customers to arrange their data-centers and design virtualized applications, you can be other vendors will support any idea that will weaken the idea (and reality) that any virtualization project must first start with products from either Microsoft or VMware.
It will confuse the comparative us-vs-them simplicity of a market in which there are two inimical, dominant players and a host of also-rans; but it will also benefit customers who will get better products, services, prices and less pointless marketing rhetoric than would be the case if every conversation about virtualization came down to which hypervisor company the customer decided to hire.