Theres been a lot of chatter about the wonderful new world of virtual Linux desktops, yet it hardly exists. Why?
When Red Hat acquired Qumranet, the company pointed to two key pieces of technology that came along with the acquisition: Qumranets hypervisor, KVM, and its desktop virtualization technology, SolidICE.
With KVM being an open-source hypervisor (its included in the Linux kernel), the real opportunity for revenue is SolidICE. Weve been hearing about the promise of virtual desktops housed in data centers for quite some time now, but how real is the market?
Theres a lot of noise around desktop virtualization, but no one has really bought into it in a big way yet, said Richard Jones, vice president and service director of data center strategies for the Burton Group, a market research firm.
Of course, considering the economy, any new technology that requires spending some money even if youll end up saving in the long run is having a tough go of it. Still, analysts are bullish on desktop virtualization.
I dont think desktop virtualization will fade, said John Madden, research director of the market-analysis firm Ovum. Its been slow to take off, but the cost benefits are compelling, and in this economic climate cost efficiency is the focus.
The benefits of virtual desktops are pretty straightforward. IT gains control and efficiency, and even maintains better security, since patches and software updates are centralized meaning IT can be confident that theyve been done.
Meanwhile, device costs can be trimmed way back, with virtual desktops being pushed out to repurposed, legacy PCs or inexpensive thin clients.
Why, then, are there few real-world examples of virtualized desktop infrastructures?
According to the research firm Ovum, part of the problem is that there is confusion about what exactly desktop virtualization means.
Hosted desktops (or thin-client computing) have been around for years, and the limitations of this model are often mistakenly attributed to desktop virtualization. With hosted desktops, individual applications are housed on a central server and pushed out to a thin client, a client that has few onboard computing capabilities.
With only applications housed centrally, not the full desktop environment, the real value is control and security.
Full desktop virtualization, on the other hand, houses the entire desktop environment on a server and delivers an experience robust enough so that end users wont even know theyre working in a virtualized environment. Multiple virtual desktops are able to run on the same physical server, yet each user has an experience consistent with full-blown PC computing, including their own individual settings.
Vendors need to do a better job of making these distinctions, of pitching real virtualization versus hosted applications.
Hosted desktops have their place usually when organizations want to limit terminals to only certain activities but too many organizations mistakenly believe that hosted desktops are as far as desktop virtualization goes.
There Are Still Plenty of Servers Left to Virtualize
The truth is that were not that far along with server virtualization, said Scott Crenshaw, Red Hats VP of the Platform Business Unit. Id estimate that we still have 90% of the worlds servers left to virtualize.
Crenshaw argues that most of todays x86 server virtualization has been with relatively simple workloads, such as Web and print servers.
The point is that there is plenty of work to be done with servers before we start worrying too much about desktops especially for organizations that dont have any near-term plans for hardware and PC upgrades.
With the economy in a tailspin, most businesses are delaying, if not freezing, new equipment purchases.
And server virtualization itself has obstacles. According to the Burton Group, the two critical obstacles impeding wider server virtualization adoption are vendor licenses and vendor support.
Many application vendors refuse to provide support if their applications are running inside virtual machines, while others support only a specific platform, usually VMwares ESX Server. Meanwhile, many end-user license agreements (EULAs) are simply out of date and dont mesh well with evolving virtualized infrastructures.
These hurdles are easy enough to clear, but will take some time.
Adoption: Slow but Steady
John Madden at Ovum, meanwhile, believes that successful server virtualization efforts will drive the demand for other types of virtualization, including storage and desktop virtualization. As IT managers realize costs savings and management efficiencies, theyll seek to replicate those successes.
I never expected desktop virtualization to have a hockey stick-adoption rate, Madden said. Adoption will be slow and steady. Its already underway, but I doubt there will be a huge flashpoint.
Both of the analysts I spoke to believe that the biggest barrier to adoption isnt technical or even cost-related. Rather, its cultural.
People are used to having their own devices. They have a sense of ownership, Madden said. From a job standpoint, they worry that virtualized desktops will affect productivity.
Another issue is support. With the desktop at a remove from the employee, it feels as if support is even more remote than before.
Its purely psychological, of course, but if those attitudes are in the boardroom and not just the cubicle, adoption is slowed.
Madden also believes the cultural issue is largely generational. Younger workers fresh out of college dont have the same attachment to PCs that older workers do. In fact, for younger workers, the go-to device isnt the PC but the mobile phone. Younger workers, though, arent the ones calling the shots when it comes to virtualization initiatives.
Meanwhile, desktop virtualization vendors are already looking past the PC to other devices. As VMware puts it, they seek to leverage virtualized infrastructures to deliver services to universal clients.
Red Hat, through its acquisition of Qumranet, has a similar vision. They believe that desktop virtualization has the potential to deliver anywhere, anytime, any-device access to computing resources.
The anywhere, any-device concept has been around for years. Remember Internet appliances?
The reality has been less exciting than the hype. Could virtualization, though, be what finally makes this vision real?
It certainly addresses a key hurdle, removing processing burdens from the client-side devices. Moreover, it promises to deliver a consistent user experience, regardless of the device.
Now, we just need better input options, better throughput over the wireless WAN and better battery life, and well be in business.