|Virtual Desktops - a blind alley?|
|Written by Administrator|
|Friday, 11 February 2011|
There are probably three hot topics in IT at the moment, "the cloud", "mobile" and "virtualisation". There is overlap between these areas but it is virtualisation I want to address in this post. Virtualisation is technology which appears to make something out of nothing. Where you might have had five physical file servers, for example, you now only need one which can be split up using software into a number of virtual servers. These in logical terms appear to be physically separate (if you browse the network from a pc you'll still see apparently x different file servers). Virtualisation at the server end makes perfect sense to me. Done right it offers huge benefits in saving on hardware costs, operational reliability, scalability and business continuity.
What I am not convinced of is the benefit of desktop virtualisation, or more specifically Virtual Desktop Infrastructure (VDI). This is a scenario where the end user, instead of having their own relatively high power pc with all its peripherals and applications, is given effectively a keyboard and screen which gives a view of an instance of the operating system or application running on a server in a remote datacentre somewhere. A 21st century equivalent of the dumb terminal "green screens" that were appearing in business up to the mid 80's, except this technology allows on demand provisioning of applications and data on pretty much any device.
The arguments for virtualised desktop infrastructure is that it is easier to manage and deploy, cheaper to install and more secure with no data saved locally than a conventional client server arrangement. The cost issue would have to be argued on a case by case basis, all I can say is that pc's are relatively cheap now and you'll still need a screen which is probably half the cost of a pc anyway, but virtualisation pushes a lot of the hardware and software costs back to the datacentre along with a stack of complexity. There would be a saving on power, particularly the cost of power distribution to desktops which in a large estate would be considerable - but that's it. To be fair Microsoft themselves don't make any claims for cost saving.
Yes, virtualising the desktops makes them easier to manage than Windows clients, but there have been tools available to configure and lock down desktops and remotely install applications available for ten years or more. Similarly "self-service" user provisioning has been possible for nearly as long. The advantage VDI has over these more traditional techniques is that it separates the application from the device and it's operating system
So what's driving this? In my view it's the legacy hardware and software manufacturers who have made their fortunes out of the dominance of the Wintel (Windows and Intel) alliance over the past decade. Windows still dominates the desktop, certainly in business, but it is almost certain that the computing device most people use on a daily basis isn't running Windows - it's their mobile, and Windows has a declining share of that space. The applications those users are accessing via their mobiles are also not likely to be Windows based either - in most cases they will be accessing the web which runs on open source technology. Facebook for example uses open source technology (see here), as does Google. Most of the web runs on open source software. For example in 2010 Apache, the open source web server, took it's share from 46.6% to 59.4%, a gain of 12.8% while Microsoft's IIS went from 21.0% to 22.2%, a 1.2% gain (source Netcraft).
The corporate world is still largely wedded to Microsoft but increasingly organisations are using a mix of technology. Apple computers are starting to appear on desktops beyond their traditional media stronghold and open source is appearing in the datacentre / server room.
Now don't get me wrong, I am not predicting the demise of Microsoft (not just yet anyway) but the once dominant corporation has got a battle on to keep up with the pace of change and user expectations.
So virtualised desktops offer advantages in some situations, but could it be that the major manufacturers are just trying to squeeze a few billion dollars more out of the Microsoft stack before user pressure to consume true web based applications (as opposed to thick client apps presented in a browser) on mobile devices finally takes over the corporate space? Will we still be talking about VDI in ten years time?