VDI - Redo or Breakthrough?

By Bernd Harzog
published: Tuesday, June 10 2008

Question - Is VDI going to break significant new ground in terms of the centralized management of end user computing, or is it simply going to re-plow the ground that Citrix has already plowed with Citrix Presentation Server (CPS) (now called XenApp)? Citrix has over 200,000 customers of CPS/XenApp but has never managed to become "THE" way to deploy all applications to all users.  So, the questions are, 1) what held Citrix back, and 2) are these same things going to hold VDI back? In my mind the key things that held Citrix back, and that VMware VDI and Citrix Xen Desktop need to overcome are:

1. Offline Operation. CPS could never support this because applications never executed on local PC's or laptops. VMware is promising a version of its VDI offering that supports both centralized and offline operation.
2. Customized User Experience. With CPS, the most efficient way to run it was to use mandatory profiles that limited users to centrally defined desktops and applications settings. This was unacceptable for most office or professional users. Roaming profiles which allowed customization, introduced management challenges and performance issues into the environment. In order for VDI to become ubiquitous, it must become possible for IT to get the benefits of centralized management, while offering the users that need it the ability to customize their environment. RTO Software has an excellent profile management solution that addresses this need.
3. Management of Applications Performance. Citrix bought Reflectent to address this issue, but the architecture of CPS itself made measuring true end user experience virtually impossible. VDI does not impose the same architecture upon all applications, so a VDI aware approach to managing applications performance is in fact possible. It looks like B-hive is not going to address the VDI  APM issues, so users will have to turn to a third party solution like vmSight.
4. Scalability. CPS certainly has the best (by far) scalability of any of the terminal server oriented solutions. However, it never could scale to meet the needs of the most demanding applications. Right now the server density numbers for VDI are not any better than the numbers for CPS. A lot of work remains to be done here.

VDI also introduces one important new issue which is how much back end storage is required to support the environment. If you have to store a complete copy of every user's operating system and set of applications then you can easily end up needing a new SAN to support the VDI environment. If you try to have a single copy of the OS and the applications then you run afoul of the customization issue discussed above.

In summary, VDI might in fact be the next greatest thing after server virtualization itself. But some significant maturation in how it is deployed and managed will be necessary before enterprises are going to be able to use it as THE mechanism to deploy all applications to all users. I think 2009 is the year in which these things will shake out. Let's see what happens.

Citrix , XenApp , Xen Desktop , 10 Minutes to Xen