John Dix, Network World Editor in Chief, set up one of his debates with industry experts. This time the question was:
Desktop virtualization: niche solution, or new desktop standard?
Desktop virtualization promises to simplify management, increase efficiency and improve security, but it doesn’t come cheap and requires a big rethink. Will the latter forever relegate desktop virtualization to a few corners of the enterprise, or will the allure of the former convince corporations it is time to start from scratch?
And on the two sides of this debate:
John Fanelli, Vice President of Enterprise Desktops and Apps at Citrix claims:
Desktop virtualization benefits drive standardization (his full opinion here)
Brian Duckering (me) claims:
Virtual desktops are a niche technology (full opinion below)
Virtual desktop technology appeals to IT because it promises to address IT problems. But most of those promises come up short and, worse, until the tech provides a positive advance for users, they will reject it. Taken together there are simply too many reasons for most people to avoid virtual desktops for it to be anything but niche for now.
What is sometimes forgotten is that IT is really all about enabling the end user. For any organization, value is created when a user is able to access computing resources and be productive. Unfortunately, desktop virtualization often forces compromise, with the strongest arguments for virtualizing the desktop coming from the IT community.
IT likes desktop virtualization because it’s all about centralizing and standardizing to simplify the management of these systems, which is often in opposition to end-user demand for greater mobility and flexibility. So, infrastructure decisions involving desktop virtualization must be done intelligently, carefully weighing the cost and convenience to IT with end-users’ need to be productive anywhere or have dedicated and responsive computing power.
We must recognize that today end users hold a lot more influence over computing than at any time in the past, insisting on flexible work locations, personally owned devices, disconnected use, etc. And in the majority of these cases, there are legitimate business drivers supporting the decisions. It is largely up to IT to figure out how to allow these new models, without compromising security or manageability.
There are certainly environments where virtual desktops are the best solution, even if the user experience is not improved, like when data must be confined to the data center, or in highly regulated industries or even offshore engineering. User productivity and mobility may be compromised, but other factors weigh more heavily. This is a niche where IT requirements actually trump user preference.
One niche where users may actually embrace virtual desktops comes from the increasing popularity of tablets. From a tablet, you can access a virtual desktop while on the move, and run all of your business applications that won’t run on the tablet OS, without having to carry a full PC. This case is becoming more popular as RDP for mobile devices continues to improve and become more prevalent. And in fact this use case may eventually move virtual desktops out of the niche category – but not until a host of other obstacles are overcome. These obstacles include:
* With virtual desktops storage costs increase dramatically as several gigabytes are transferred from cheap distributed storage to expensive data center storage for every converted desktop
* License management is challenging in any environment, but virtual desktops make it even harder. Traditional management systems weren’t designed for this dynamic environment, and often more licenses are required to ensure compliance.
* Infrastructure costs can be enormous, with storage being only one factor. Add thin client devices, VDI licensing, various servers, plus additional OS licenses in the form of VECD or VDA. This is rarely a break-even financial decision, with CAPEX often outweighing any predicted OPEX savings. Financial arguments will rarely point to this as a solution.
* Maintenance costs are one of the biggest promises of virtual desktops. But reducing operating costs is really only possible at near 100% saturation. At 15% – 25% (the high end of most practical implementations), the existing landscape does not change much and these new costs become additive. Plus, everyone accessing a virtual desktop must do so from a physical one, potentially doubling the number of endpoints to be managed.
How long virtual desktops remain niche will depend on how long it takes to solve these problems and how long it will take to convince users they will be better off. In the meantime, there are other good alternatives, depending on the objective.
If it is just a matter of centralizing computing, the terminal server approach works well for most, and is far less expensive. For better management of distributed systems, streaming is a mature and proven technology that centralizes management, automates user customization, improves delivery and removal of apps, and reduces helpdesk calls, among other things.
There is no one best solution for everybody today, and IT managers should be encouraged to pick the best model for each user group. Heterogeneity will be with us for a long time.