A couple of years ago, UIAs (user-installed applications) were all the rage. AppSense had a product called StrataApps (and you can do it in Application Manager also), RES has native functionality that allows this, Unidesk can do it (admittedly in a slightly different way with its layering technology), Citrix was pushing Personal vDisk with its acquisition of RingCube—the list goes on. However, the actual use of UIA functionality has been limited, to say the least, and the response to it from IT departments remains frosty at best. Despite predictions that UIA requirements would become an integral part of desktop and (to a lesser extent) server virtualization projects, they seem to have quietly faded away. What happened to the UIAs?
The Premise of UIAs
Many will have seen this diagram demonstrating the “longtail” applications issue before:
At the start, you have applications that are used by lots of users lots of the time—Office, SAP, Acrobat Reader, browsers, etc. These applications are easily dealt with—installed into base images and kept up to date there.
In the middle, there are the applications that are used by a subset of users lots of the time, or by all users infrequently. Think departmental applications, MS Visio/Project, AutoCAD, etc. These applications are prime candidates for virtualization through technologies such as App-V, Citrix, and ThinApp—delivered seamlessly via a client (or even through the browser directly) without any need to perform complex installation routines into base images. They can even be packaged online, if you can integrate them there—think solutions such as Spoon.net.
At the end, you have what are referred to as the “longtail apps.” These are applications used by very small subsets of users, often very infrequently—possibly only once a year or so. Things I’ve seen in the longtail are apps like TweetDeck, Notepad++, CD-burning software, and Lotus Notes (where Notes was no longer the email client).
The idea behind this was that the longtail apps would not require publishing, packaging, installing, or managing in any way—ideal for resource-starved IT departments where there is no sense in spending days configuring an application that a user might use only once a year. Via the UIA technology of choice, users could simply install the longtail apps as they needed them, without having to contact support or gain elevated privileges to do so. Indeed, the idea was quite valid, and the solutions were in many cases quite good.
There are a few possibilities as to why UIA technologies simply haven’t taken off:
It seems certain that many IT departments are resistant to the idea of UIAs because they don’t trust their users and are reluctant to extend any sort of privilege to them. Despite everything we read about how the workforce is now tech-savvy, mobile, and fluid, a great number of IT departments still insist on providing a monolithic desktop with no flexibility whatsoever. Even with solutions that provide pre-approval mechanisms for UIAs, IT departments appear paranoid about ways in which extending trust to their users could be abused. To most of us out there, it appears monumentally old fashioned of IT to behave in this way—but it has to be said that many companies still act in traditional manners.
Another worry is that allowing users to install applications is going to prove confusing for them. This appears even less valid than the trust argument, as users are now accustomed to installing applications on home computers, tablets, and smartphones. While it remains true that some users may not want this functionality for fear of the complexity, it does not need to be extended to everyone. Indeed, one of the prime drivers for UIA solutions was to assist expert or “super” users in maintaining their own environments.
There’s also the possibility that users aren’t that bothered about not having access to other applications anymore because they have them installed on their laptop/tablet/phone and can use that at work. Certainly that seems a possibility; as I type this, I have my own laptop next to me, joined to the corporate wireless network, with all of my applications installed on it. But there are many places where security is not as permissive. Although some of these applications can be used offline, sharing data is probably not as easy in secured networks, where USB drives are restricted and websites like Dropbox are blocked.
Possibly the most compelling reason for the apparent failure of UIA tech is that advances in other areas have made it unnecessary. Certainly it is now possible to provide a persistent, one-to-one hosted client desktop for a reasonable price, mainly due to advances in storage solutions. This sort of advance in turn makes the whole “profile” aspect of the user less necessary, not just the application sets. Layering tech, although sometimes rather complex, also adds to this. Have we simply reached a stage where UIAs are unnecessary?
Are There Any UIAs Out There?
Of course, there’s also the distinct possibility that my experiences are not indicative of the true state of affairs, and that UIA deployment is alive and well in the enterprise today. While the wide variety of solutions that either offer or encompass UIAs may suggest this, there’s also the possibility that they’ve all embraced this not because of a need for it, but because their competitors have it as a feature and they want to claim to offer that feature as well.
So, are UIAs alive or dead? They certainly sprang up offering to fill a need in the enterprise, but somehow, they seem to have either faded away or been implemented without much fanfare. Have people simply used technology to make the need for a UIA solution less pressing? Or are the users simply not requesting the feature due to managing the applications themselves? This possibility seems the most worrying, as it would mean that application (and possibly data) management has been removed from the governance of IT—a bad situation for many businesses to find themselves in.
Share this Article:
Latest posts by James Rankin (see all)
- Anatomy of a Desktop Virtualization Project #3: PoCs and Pilots - July 8, 2015
- Can One OS Rule Them All? - April 29, 2015
- In Search of the “Nirvana App”: Cloudhouse, Next-Generation Application Virtualization - March 19, 2015