I started working with Virtualization when VMware made it available to the x86 environment, little more than a decade ago. People didn’t really know about it and things that we take for granted these days, were quite difficult to do. For that reason, take up was a bit slow but in recent years with technologies maturing, it has really taken off – to the extent that it has even become fashionable! (Imagine going to a party with no virtualization in your datacentre – expect difficult questions!) Seriously though, the situation is also helped by the economic situation as companies try to cut costs. That said, while there still are a lot of companies that need to convert, many have already virtualized their server environments and are now turning their attention towards VDI (Virtual Desktop Infrastructure).
Apart from the obvious benefits of virtualization, it also means less support and maintenance costs, desktop in the datacentre (read “security”) and easier rollouts or rebuilds etc. But is this the right time to invest in the technology? My view is that companies should at least start experimenting with it. With Application Virtualization also becoming mainstream, there was never a better time to give serious thought to desktop strategy for the next few years. It’s true that current hardware-based PCoIP clients are not cheap so ROI is not clear and will take some time to materialize but there is no reason why companies can’t start with a software-based client on their existing desktops today. The trick is to get the architecture nailed down while these technologies mature and prices come down. A gradual shift to this architecture will not only give a chance to iron out any issues and get staff to become comfortable with it but will also not tie the whole environment into the technology while it’s still new.
Moreover, none of the VDI implementations that exist today can survive the loss of network connectivity. That coupled with the fact that virtual desktops have limitations in certain areas e.g. multimedia, makes it a difficult sell. Convincing the user is absolutely crucial when introducing new technologies and if they’re not able to do their job properly as a result of the change, the project is likely to fail. I am sure all vendors are working on their solutions for these weaknesses but until these problems are resolved, virtual desktops will have to coexist with traditional ones. This enforces a requirement on the architect to ensure everything works seamlessly for a particular user, regardless of the type of client. Slowly introducing such desktops into the environment will help prepare for a seamless rollout when it’s actually required.
We know by experience that not every server is a candidate for virtualization. In the same way, there are cases where VDI is not appropriate. However, if there is a large desktop population and replacement time is coming, maybe it’s an opportunity to replace the desktops with a VDI-equivalent. Like everything else, moderation is key here. Developing a long-term strategy with a mixture of both types of client and an aim to virtualize as many applications as possible must surely be the way to go.
… Now if only I could get my hands on that VDI Starter Pack 🙂