2011 Prediction: MokaFive
Virtualization has come a long way over the past decade. Ten years ago, VMware had recently spun out of Stanford and released its first product, VMware Workstation. Now, virtualization is a multi-billion-dollar industry and is in use in nearly every large enterprise worldwide. Virtualization has evolved from being used for testing and sandboxing in the early 2000s, to server consolidation mid-decade, to most recently cloud computing and desktop virtualization.
However, the widespread adoption of desktop virtualization has moved forward at a much slower pace than server virtualization, and there are a number of factors that play into this including the complexity of the problems it solves and the fact it touches many more people. Perhaps an even larger factor stems from most vendors still viewing desktop virtualization through the lens of server virtualization. They often try to solve the desktop problem the same way they solved the server problem – namely by consolidating resources into the datacenter – in turn shortchanging the problems of how to effectively manage the desktops and the need to be able to work offline. Many companies have conducted trials of server-hosted desktops (VDI), but there have been few large deployments.
Looking forward to 2011, more and more of these VDI projects will be shot in the head because the numbers just don’t add up. VDI is turning out to be a far more expensive proposition than people originally anticipated. To achieve a good experience with VDI requires a lot of server horsepower, a fast network, screaming fast SAN storage, and dedicated people to manage and maintain the system. When it comes to evaluating the return on investment, more and more companies will realize it does not make sense and restrict VDI usage to a small subset of use cases, or cancel it entirely.
An alternative and more cost-effective approach to desktop virtualization performs the computation on the endpoint client device, while maintaining centralized control. There are two predominant models for client-side virtualization: type-1 hypervisors and type-2 hypervisors. A type-1 hypervisor runs directly on the hardware, whereas a type-2 runs on top of an existing operating system. Citrix released their type-1 XenClient 1.0 client hypervisor this year, but the technology is still very primitive.
In 2011, type-1 client hypervisors still won’t achieve significant adoption. XenClient and others will continue to improve their technology and will get closer to an offering that is actually deployable, but because type-1 hypervisors are tied to the hardware, they are limited by the hardware refresh cycle, which companies are starting to extend out to as much as three to five years. Type-1 client hypervisors won’t have a significant impact until they are built into the hardware from the manufacturer, so we won’t see widespread adoption until 2014.
Type-2 hypervisors will continue to have a compelling use case in “Bring Your Own Computer” (BYOC) and work-from-home scenarios, because of the need to leverage a user’s existing hardware and the fact that wiping the machine to install a type-1 isn’t practical. The primary drivers of security, mobility, and flexibility will continue to play the largest role in the further adoption of desktop virtualization solutions in 2011.