We often make the point that since the Hypervisor market is nascent, you don’t quite know whether what is best of breed today will remain so 12 or 18 months down the road. This reality must be factored into your decision making process when you choose a VDI vendor. Down the road, when you swap out the Hypervisor for another, the rest of your VDI stack shouldn’t come tumbling down! Unfortunately, since many hypervisor companies are so completely dependent on their Hypervisor revenue streams, they are also doing everything they can to tie their Virtual Desktop solutions exclusively┬áto their brand of hypervisor. VDM2 from VMware being one example of this phenomenon.
A recent game of verbal ping-pong between Simon Crosby, Citrix’ CTO and Brian Madden, Virtualization Blogger Extraordinaire,┬áhas shed further light on just how important Hypervisor-agnosticism truly is. If you read the full posts on either side, you’ll find that there is general disagreement on how different Hypervisor platforms are going to evolve, whether key players will continue to invest in certain platforms down the road, and whether newcomers, such as KVM, have any legs to stand on in terms of market acceptance (not just technology).
It started with Brian Madden’s prediction that Citrix was soon going to abandon the Xen Hypervisor platform in favour of Hyper-V. He premised this on┬áXenServer having “zero” penetration, on the similarity of Hyper-V and Xen architecture and on the rise of KVM.
Simon Crosby responded with the observation that Xen is technologically superior to Hyper-V and Citrix has no plans of abandoning it. He suggested that Hyper-V is a “charging bull” aimed directly at VMware (thus confirming what we have shared with you earlier; VMware’s own future is uncertain). And that KVM is “late to the party” and unlikely to succeed because it doesn’t present any radical advantage over and above what’s already available on the market.
Regardless of who you agree with, the bottom line is that the Hypervisor market is in flux. The effects of Hyper-V’s RTM remain to be seen. Xen is a contender for sure, and its open nature make it perfect for large EC2 (Amazon Elastic Cloud) like implementations. There will be a lot more of these popping up, and they will power the future of cloud computing, therefore┬áXen certainly cannot be written off.┬áWhether or not Citrix continues to sell XenServer is immaterial, because the Open Source Xen project will continue to be available regardless of whether Citrix packages it in their XenServer product.
While its hard to argue with the benefits that VDI brings to business, it is also true that making the wrong bet on VDI technology can put a bad tase in your mouth and ruin your ROI calculations. If you have to migrate to a different Hypervisor 12-18 months down the road will you also need to re-wire your management and connection brokering infrastructure? Will you have to reimplement what took tens or hundreds of thousands of dollars in employee and contractor time to get right the first time? If you choose a technology like VDM, which only works with one type of Hypervisor, then the answer to all of the above is a resounding, “Yes”.
So, for customers, Hypervisor-agnosticism is KEY. And while the Hypervisor is an important part of a VDI deployment, it isn’t the only necessary infrastructure component; the remoting protocol is another. We’ll dive into that in depth at a later time and explore what implications remoting protocol choice have on the Quality of Service you can deliver to VDI users. More later.