Thin-client computing—a term to strike fear into many users, with thoughts of a weird desktop device giving slow, inconsistent access to applications, little support for local devices and no capability for untethered use. Possibly, at a stretch, useful for those who are completely "tethered" (remaining sitting at the same desk day after day) to a single desk, carrying out repeatable tasks, but surely not for knowledge workers and executives?

Quocirca has historically agreed—thin-client computing met the needs of possibly 20% of an organisation, but the constraints and restrictions meant that it was not an optimal solution for others. But technology changes, and now we can take a new look at how this market has changed, and whether more of the employee base can be moved to a position where their "desktop" is actually held in a more controllable central location.

Firstly, let's get away from this being a thin-client argument. Fine, if you want to go for a specialised Linux or proprietary operating system-based machine that supports a browser capable of supporting a server-based desktop image, then there's no reason why not, other than making sure that sound can be streamed to the device, and that local devices can be supported. However, companies such as Igel and Cranberry provide low-power, low-cost devices that are based on a standard Windows operating system, maximising the capabilities for supporting local devices, plus offering a degree of hybrid capabilities, with the device being able to carry some of the workload locally, rather than being fully dependent on the server for all computing power. And, of course, there is nothing intrinsically wrong with using a standard PC as an access device—it will take more power to run, and may have a higher propensity for failure due to the increased complexity of the device, but with desktop prices falling to similar prices for many "thin clients", best to make the most out of what assets you have available to you, replacing failed PCs with newer devices as necessary, rather than as a rip-and-replace programme.

So, why does Quocirca now believe that organisation should now be looking far more to moving the desktop to the datacentre? Centralised control, better security and information management along with greater licence management capability have always been put forward as good reasons for the move, but there has to be more. Virtualisation changes the discussion in many ways—more desktop instances can be supported on the same hardware, and provisioning new desktops is easier than it was.

There is also the capability for a server-based image to run "sandboxed" against an access device. Here, an employee can have their own device, which may be being used for private use, and as such may be prone to malware and other security issues. Although an organisation could try and lock down the device and enforce its corporate standards, the user would not be happy, and will generally find some way around such constraints. However, by presenting an image in a secure environment in a manner where nothing at the main device operating system or application level can pass any data through to the running image, a far more measured approach is provided that keeps both sides happy.

But, the real change that virtualisation is bringing to the fore are the new tools that will allow a far more granular means of managing images. Historically, if you had 1,000 users in a thin-client world, then you had 1,000 images—each with an operating system and the rest of the associated stack of applications. If any one of these needed patching or upgrading, 1,000 images had to be addressed.

Now, were beginning to see that newer approaches are providing solutions where "golden" images at a granular level are being kept. Here, for those 1,000 desktops, there may be 3 or 4 different versions of operating system required, and so only 3 or 4 images are kept. If the operating system needs patching or updating, only these golden images need updating. As a user needs to access their desktop, the complete image is built dynamically from the golden images, providing far greater management and control capabilities.

Also, the main vendors in the space have realised that a major issue with server-based computing is in the growing mobility in the workforce. Although the promise of an "always on" connection is becoming more of a reality day by day, these connections do not always provide a consistent experience, and for those looking to being able to work on a plane or in out-of-the-way places, a connection is still not a viable option.

So, part of the server-based computing offer is now centring around the capability to provide a fully synchronised experience between laptop and server for those who are highly mobile. This then enables an organisation to exercise greater control over field employees, such as engineers and sales forces, and also to support executives far more in the manner that they expect.

As part of this mobility, we also have the use of different devices—the mobile phone has become a far more intelligent device than we have seen in the past, and the emergence of devices such as netbooks bring different access and management needs to the fore. With a server-based computing approach, a lower degree of common intelligence is needed at the access device, and so greater levels of access can be provided to each user, irrespective of the device type they are using.

So, greater information security, greater license control, enhanced support for local devices, support for VoIP and video conferencing, untethered capabilities, greater optimisation through the use of virtualisation and greater end device support. Altogether a pretty good set of arguments for organisations to revisit their desktop strategies, looking to introduce a broader usage of server-based computing.

Quocirca has published a report on the subject, that can be downloaded free of charge.