With Windows 7 deployments picking up speed, there is a lot of discussion about desktop refresh and/or virtual desktop infrastructure (VDI). Servers and storage costs have come down, and the performance constraints we saw a few years earlier have been addressed. The conversations usually have two starting points cost and management. However, I usually see a third point come up towards the discovery phase that 9 out 10 times kills the VDI project, end user experience.
Cost is an easy and somewhat concrete measurement for VDI. The formula is Hardware costs x number of users = cost. If you reduce hardware costs, through repurpose or thin clients, your overall investment will be less. It isn’t too complicated and most of the time this is the driving force behind researching an alternative to a true hardware refresh. With that said, most research is showing that the average corporate desktop is now 3+ years old and even worse, won’t run Windows 7 to its full capabilities. Throw in out of warranty machines and rising help desk costs and it seems like a no brainer.
I don’t want to spend too much time on the full ROI or TCO of owning a VDI environment. It can get complicated fast, and there is a lot of “what if’s” and other soft data. And let’s not forget licensing and compliance issues involved. It can be a very long and detailed process. One thing I do want to make clear on costs is that if IT wants to show a savings, they can. The numbers can be fudged in either direction depending on the tide of the internal staff. In short, my point is that cost rarely, in my in opinion kills a VDI project.
For the management piece, VDI makes the most sense and if you can’t build a case around cost reduction, this is a great tool to build a case around. Think about how much can be done by deploying a whole office upgrade in minutes, or trying out a patch and being able to restore a snapshot if something goes wrong. Or even taking on new employees and you need to deploy 20 machines so they can start work on day 1. My favorite use so far is sending out a packaged thin app to remote works, with a note that just says “Plug in, press ENTER.” Worried about sensitive data leaving your network on a laptop, no more if you build out a best practice VDI solution; all of your data is centralized.
People get really excited as they hear about the benefits and everything goes well until the proof of concept (POC) stage. The thing is, people expect to interact with their computers in a certain way and when things change, even for the better, there is always a small revolt. Think Office 2003 to Office 2007, not so happy about the ribbon, or the change to Linux from Windows, disaster. VDI is not much different. Things tend to move slowly but things seem the same, but different at the same time. Web pages lag behind; applications don’t open, missing items on the desktop, etc. If you are working with a more technical crowd there is a level of paranoia that creeps in, a big brother type of mentality. Is IT spying on me? Where is my old machine?
This ultimately leads to more helpdesk tickets, slower adoption and morale regression. However, I am not saying that VDI is bad, far from it; my objective here is to rethink the VDI paradigm and address it a little differently. I would suggest starting from the “user experience” side and working backwards. There are a lot of companies out there, Citrix, VMware, Microsoft, RedHat, Oracle, etc that are offering new and better solutions all the time. Things are improving and we are seeing creative and imaginative solutions to some pretty complex issues. Below are some starting questions I would ask to myself and team before jumping with both feet.
- Can our current infrastructure truly support a complete VDI environment?
- What assessment tool are we using to get these statistics?
- Do we have the resources to build, test, and rebuild based on our corporate need?
- What is our alternative if an end user doesn’t wish to use VDI (mixed environment, no exception policy, etc.)?
- For a POC what qualifications do we need to hit? Performance? Cost Per Machine?
- What programs, and uses will this be used for, will poor performance hinder or stop my current production process? Will this add or remove our corporate efficiency?
Once you have addressed these questions I would also recommend building out a long term plan and match that to the vendor of your choice. While I will go into reviews in future posts, I will just say briefly it is extremely important to understand your VDI vendor’s roadmap. You don’t want to pick a company that is either half invested in VDI or is planning to move in another direction and be stuck with an obsolete product.