VDI- One Man’s Trash Is Another Man’s Treasure, Or, Why Crap Is King….

August 9, 2011

[Please note, is under renovation- some content and links are in still in progress]

I.T. Professionals and Consultants who have worked for any period of time on hosting (or virtualizing) applications and desktops are acutely aware of the unstructured data that becomes part of a user’s environment. On a standalone PC it goes pretty much unnoticed as it “blends into the woodwork” of the overall system, spreading itself across the registry, file system and user profile. However, when you virtualize applications and desktops you become faced with trying to capture and re-apply this data as users move across diverse systems. Tim Mangan identified this issue in his 2008 Briforum Session “The Data Problem” which was an early recognition of the problem and a great explanation of the sources and impacts (PS-that’s the back of my bald head in the audience). He also has a more recent article on the subject  “How to Describe Layering: the blob, cake, or 3D Tetris”.

Over many years of working with Roy Tokeshi, a leading Citrix SE,  he would refer to this set of data in his technical/business presentations as “Crap”. In an effort to validate this concept, and to be able to actually use the word “Crap” in presentations, I came up with the following acronyon:

Computer Residue of Applications and Personalization (C.R.A.P)

I was pretty proud of this one and then Ron Oglesby pointed out on Twitter that “I love your acronym. But Users are like Hoarders. Some guy’s CRAP is their meaningful “stuff’ ”  

As a result I am releasing an alternate version:

Carefully Retained Applications and Personalization (C.R.A.P)

So now we can use “Crap” in any context , positive or negative, to refer to this same set of undefined data that attaches itself to users and applications.

This a strange problem because on the one hand our inclination is to simply retain all this data and carry it across whatever environment the user wants to run in. Whenever possible we like to have the settings that a user expects automagically appear (because then people are happy and we are heroes). Yet, large portions of this data may be  irrelevant (at best) or even incompatible (at worst). This problems shows itself most acutely in mixed environments where applications are delivered across multiple operating systems, and, when using other tools such as App-V. For example, a user may have a local desktop OS (i.e. XP), a hosted VDI desktop OS (Win7) and apps or desktops hosted in Windows 2003 and 2008 R2. In these cases there will be corruptions of settings, locked sessions, broken profiles, etc. when indiscriminately mixing this data across platforms.

What is the solution? Well there is no simple answer that can be applied in all cases, but it comes down to knowing your applications and including/excluding the correct portions of the data for the target platform. The details will follow in a future entry, but for now we have identified and understand the challenge this presents….

Let's discuss a strategic IT partnership today!