The cost of true VDI

By Mike Jansen, DinamiQs Lead Architect for VirtualStorm

Virtual Desktop Infrastructure (VDI) has always held promise. In the drive towards centralized control over IT environments within corporations, VDI is now a subject that is always on the agenda, somewhere at the bottom, which means it is rarely addressed. In part this is due to complexity of the subject. After all, desktops are the tools that employees use to generate the work that drives the profit of these corporations. Changing anything in that environment requires careful deliberation and compelling arguments. So far, most arguments for VDI center on efficiency. And that is a soft premise to build a case on. Who can measure efficiency and in such a way that it actually makes sense?

Right now, with actual full-blown PCs on people's desks, the economic model is clear. You buy a box, you put your applications on it, you keep your fingers crossed that all installs well and without glitches, preferably from a single source image. Then in three or four year's time the box becomes slow, the technology obsolete and worthless and the whole cycle restarts. In the mean time you have wasted valuable resources on keeping the environment up-to-date and the user's desktop as clean and effortless as possible. Unfortunately the true cost for fat client desktops is high, but it is the only thing that people know, it's very quantifiable and therefore it's controlled and understandable. But it in this day and age looking at alternatives is not just economic but also necessary. There is a competitive edge to be gained through new technology, after all, innovation loves a crisis and VDI is such an innovation.

In order to get VDI back to the top of the list, it will have to become so interesting, so compelling that any company not talking about it is sure to lose its competitive edge. But as with all new technologies, understanding the implications of doing things differently and the complexity of existing environments make deciding on the proper solution for your business a real challenge. Add to that the hype that is being created around the subject and the many vendors who try to sell you their product as the end-all solution for desktop virtualization and the decision on what is best for your enterprise becomes a difficult one. Change is never easy. Or is it?

In fact, the move towards a centralized desktop environment is nothing new. We were doing it forty years ago on mainframes, where each user would be given access to their own piece of workspace and where there was full control over the centralized environment. VDI is basically a move back in time, to the age of the mainframe. No more isolated desktops that require extensive and complex maintenance and that have to be upgraded and updated in an everlasting cycle of new hardware and software. The mainframes were very successful in their day, but the costs associated with these environments were such that it made sense to acquire smaller devices for specific desktops within the IT landscape. The efficiency and speed of this model lead to the wild spread of desktop machines and the decline of the centralized mainframe for everyday work. It also lead to a client-server model and the decentralized management model for maintenance and support of the environment. And although it initially was more cost-effective to run single desktops for specific tasks, scaling out this model turned out to be as expensive or even more so than using mainframe style environments. Most of this caused by the huge increase in maintenance and support of the environment, although on the CAPEX front the investments in new and more powerful hardware declined every year thanks to Moore's Law. But buying more and ever cheaper hardware will not solve your problems: it will only increase them. All this new hardware requires software and licenses and support in addition to what existed before. Face it; our current way of handling IT environments is unsustainable from a managerial and support point of view.

For VDI to actually make sense it should simplify, streamline and solve problems instead of creating them. It should make use of available resources and should be optimized for all workloads. It should be cost-effective from a CAPEX standpoint, but it should also drastically decrease your TCO (easy in full desktop environment with many applications) and finally, it should start paying for itself within months instead of within years. The CAPEX discussion is easy; hardware gets cheaper every year. But TCO tends to increase as more hardware is dragged into the environment. And ROI is a moving target because it's hard to quantify gains and losses for any investment. So the best case for VDI would be to decrease the actual TCO of running the entire IT environment. But for that to happen, VDI would have to become a huge cost saving compared not only for the desktop, but also for the overall IT environment.

So, how to get TCO for VDI to make sense and actually make it the top priority for the CxOs and IT managers? The answer is actually easy: change exactly nothing (for the end-user) while going back to a mainframe style maintenance and support: centralized, predictable and scalable. That may sound too easy, even too good to be true. But the technology exists today in DinamiQs' VirtualStorm, a True VDI solution that uses extreme virtualization of the entire desktop hardware and application stack. It's time to get back to the mainframe, but this time it's a desktop mainframe and the big winner is the end-user and therefore, ultimately, the enterprise.

For more information: www.VirtualStorm.org