One major misstep made with VDI is not recognizing that the storage workload for a desktop is very different from that of a server. From a technical standpoint, VDI is a lot more balanced for read and writes than a typical server. Often, server workloads are tuned for fast reads, while virtual desktop technology tends to be split more down the middle. What happens if you use your typical datacenter storage -- tuned for reads -- for VDI? According to Madden, it won't necessarily work the way you want it to.
Latency is a key issue for all projects, but it presents a distinct problem in a desktop environment. If there is latency on a back-end server, a user probably won't notice. However, if someone clicks the Start menu or opens a Word document and it doesn't pop up instantly, the helpdesk will get some calls.
The final unique requirement of virtual desktop technology is to plan for the unpredictability of users. With storage planning for servers, you can strategize for busy times. But as Madden pointed out, desktop users can go from creating a document in Word to opening an email, to clicking on a link in that email to watching a video in just a couple of minutes. The storage needs to be able to cope with rapidly changing workloads.
"We have to understand that the storage we build for VDI has to be flexible enough to support users doing whatever they want at a moment's notice, and it has to deliver them a good experience," Madden explained.