February 21, 2011, 2:42 PM — The most telling sentence in Michael Fox' DeMystifying the Virtual Desktop has nothing to do with the complexities of the technology itself, the endless variety of options or competition among similar but largely incompatible implementations.
It explains why virtual desktops have been among the hottest topics in IT for the last half decade, but haven't ever seemed to hit in the way other technologies have:
"Desktop virtualization is not a blatant technology. It may never be. It is a game-changing technology."
It sounds contradictory to say a game-changer of any kind has no blatant impact. It's not.
Desktop virtualization is not an independent technology in itself, as virtual servers, cloud computing, storage networking or other product categories are.
It's really an access method implemented to let a specific class of end users do specific things without blowing the whole IT budget or exposing it to too much risk.
Right now that usually means giving highly mobile end users full access to their data and applications from smartphones, tablets or their own home computers, using the same secure, standardized "desktop" environments that would exist on a company-issued computer used inside the firewall.
A few years ago it meant letting a call center full of hourly workers access the same application and data on a single server and single OS to minimize cost and maximize availability and data retention.
Either way, the generic image most people have of "desktop virtualization" covers one thin slice of a very large pie.
Large-scale desktop virtualization projects typically involve three to five major types or delivery mechanisms for a "virtual desktop" ranging from a single application streamed on demand to a desktop, laptop or mobile device, to full-scale VDI – discrete virtual machines that include the operating system, applications, utilities and all the other accoutrements of a real PC dedicated to a single user but running on a back-end server rather than a local piece of hardware.
As a result, "desktop virtualization" decisions tend to be limited in scope or implemented in pieces, user by user, department by department. Migrations, even those launched by top-down strategy decisions, don't take the place by storm. They seep through it like leaky pipes hidden in the walls.
Deciding which users should get what resources using what mix of virtual-delivery mechanisms – not to mention the network, storage and server resources to support them – is among the more complex problems in IT.
"DeMystifying the Virtual Desktop" can't make questions simpler, but does make understanding and answering them easier by laying out not only the technologies and options available, but by describing the process of identifying types of users and the kinds of delivery method most appropriate for each.
Laying out the territory and options is useful. Laying out the process of assessing what you already have and how to lay out a virtual-desktop architecture that will support your user base without collapsing your existing infrastructure are the most valuable contributions.
Detailed analysis steps and thumbnail capacity measures help walk you through detailed assessments of user requirements and existing capabilities.
Design analysis that includes not only steps in the process, but documents you'll need, models you should build, and analyses that will help you define the architecture and defend the spending for it are the closest I've seen to a direct, usable how-to guide for virtual-desktop projects.
There are plenty of guides to desktop virtualization out there. Most are written by and for particular vendors.
DeMystifying author Michael Fox is an experienced consultant who currently works as a senior architect for the professional services group at EMC, parent company of virtual-desktop player VMware.
Nevertheless, his guide takes a neutral approach to the technology and decisions surrounding it. He offers relatively objective guidance on a process that is notoriously idiosyncratic, complex and risky, because every step has a direct effect on the satisfaction and performance of end users.
Kevin Fogarty writes about enterprise IT for ITworld. Follow him on Twitter @KevinFogarty.
