Most large organisations have already successfully deployed a level of server virtualisation. The benefits are clear and immediate, mainly around improved total cost of ownership as a result of consolidation and better utilisation.
Organisations should continue to channel reasonable resources towards server virtualisation. It is estimated that 20 to 25% of installed servers have been virtualised in 2010, which leaves considerable room for additional efficiency gains, writes Sean Wainer, country manager at Citrix Systems SA.
Although there is plenty of scope left for savings within server virtualisation, it is, of course, entirely correct that organisations should investigate other ways of improving IT efficiencies.
One of the areas of increasing interest is desktop virtualisation, which decouples hardware and software at the end-user level; primarily desktop PCs and laptops, although it can also include netbooks, tablets, smartphones and other mobile computing devices.
Desktop virtualisation delivers considerable efficiencies in the management of end-user devices – especially across a large estate where thousands of desktops can be centrally managed – as well as improved security and subtler benefits such as extending the life of older PCs and supporting new ways of working; for instance, working from home which in turn eases demand for corporate office space.
Like server virtualisation, the bulk of the efficiencies to be gained through desktop virtualisation stems from centralised management. However, to be successful, desktop virtualisation demands different – and a much broader range of – technology. Less obvious, but of more importance, it also requires a very different strategic approach.
Having already achieved some success with server virtualisation, many organisations, often keenly encouraged by its server virtualisation vendor, take the decision to implement desktop virtualisation. That very point of inception, in itself, demonstrates the flawed thinking that leads to desktop virtualisation failure.
The assumption that desktop virtualisation can be pretty much a reiteration of server virtualisation is wrong. For a start, unlike server virtualisation, it cannot be implemented through a unilateral decision nor implemented centrally without anyone beyond the IT department knowing or caring.
The IT department and end-users have different agendas. End-users want to be able to access streaming video and audio. The IT department wants to cut budget and network bandwidth. The danger is that, with a server virtualisation approach, the IT department only sees that the management of 100 PCs is being reduced to the management of three servers.
In doing so, however, IT managers forget what that means to end-users; both the quality of the user experience and the concern end-users will have about their computing environment being delivered differently.
The internal team that ran the server virtualisation team, particularly at leadership and corporate sponsor levels, may not be the right people to head a project that requires a great deal of interaction with end-users whose "buy-in" is essential.
Likewise, because desktop virtualisation involves a very different combination of technologies, the incumbent server virtualisation vendor and/or services firm is not automatically suited to desktop virtualisation. An organisation should treat desktop virtualisation as a brand new challenge, and in doing so thoroughly assess the desktop virtualisation credentials and proof points for any proposed vendor and services firm.
In short, there is very little relationship between server and desktop virtualisation. A successful initiative in one, be it proof-of-concept or cross-enterprise, does not indicate a similar outcome in the other.
Largely contained within the data centre, server virtualisation is an initiative that has a feast of low-hanging fruit. The domain of the IT department, the strategy and technologies deployed inside the data centre are made largely independent of end-users. The vast majority of servers can be readily managed and maintained in a uniform manner, with the only visible outputs being self-manicured efficiency gains.
Rows and rows of machines, doing the same thing from the same place, day-in, day-out, and subject to a consistent command and control structure that completely respects and abides by hierarchy. To apply the military discipline of the server virtualisation approach to managing employees’ PCs is to condemn desktop virtualisation to failure.
The server virtualisation approach dictates a uniform solution – for desktop virtualisation that is generally virtual desktop infrastructure (VDI) – and seeks to impose that solution on as many PCs as possible.
Now remember that server virtualisation has only reached around 20 to 25% of all servers, and that servers are reasonably homogenous. The 80% of servers that are yet to be virtualised require a little more finesse than a generic "one size fits all" solution.
By sticking doggedly to VDI for desktop virtualisation, the rigid world of server virtualisation is approaching the considerable differences between the ways people work with just one technology. That is akin to a football team always playing to the same tactics regardless of the opposition.
In an age where extremely powerful, ultra sleek and easy to use computers are being widely used by consumers, the mere notion of a workplace delivering its employees "any type of PC so long as it’s black" is completely outmoded.
While server virtualisation is about consolidation and begins at the data centre, desktop virtualisation is about making desktop management more efficient and begins with end-users.
The end-user computing needs of an office-based task worker, a field engineer and a senior executive who travels a lot differ significantly. They will have very different computing devices for a start, let alone types of applications, security needs and styles of working.
Each user group, particularly those who need their computer to deliver quickly and reliably each and every time, will flinch at the thought of change and fight against it. Not that they are likely to be 100% satisfied with their current IT set up, but they can certainly imagine the disruption that change can bring, and are likely to have been sorely disappointed by promised improvements in the past that never arrived.
A genuine desktop virtualisation approach therefore has to encompass a range of technologies, ranging from server-side computing (such as VDI) where the data centre delivers the computing environment to client-side computing that supports offline working in a virtualised environment.
A portfolio of desktop virtualisation technologies means the IT department can work with end-users to assess how they work, so that each employee group can be supported by the most appropriate technology. Individual end-users, therefore, enjoy an improved computing experience while the organisation benefits from the centralised control and management of desktops.
At a desktop level, user experience is key. Any project that does not maintain or improve the user experience will hit major resistance. In this context, people don’t necessarily want cutting edge, however they always want the right information in the quickest way possible.
That is a very different mindset to 20 years ago when employees used to compare the PC processor speed of their computer as a way to measure themselves against peers.
As an example, most organisations need to be able to access rich media (such as Flash video) but without creating excessive bandwidth demand.
Within a specific campus that is relatively straight forward, but including all types of employees – including provision across wide area or virtual private networks for mobile and home workers – means the delivery of desktop virtualisation needs to support rich media across a latent, low bandwidth network without a discernable drop off in performance.
That is a perfectly deliverable through desktop virtualisation, but the application of “caveat emptor” (let the buyer beware) is crucial as sharp practice and slick marketing can mask the shortcomings in many vendors’ offerings.
Too often in technology sales, the dream is sold over the reality, so the end-user organisation should pay particular attention to these types of detail and ask to see customer sites where the solution is already in operation.
Whatever the perceived cost saving from an IT perspective, a flawed desktop virtualisation initiative that results in employees being less productive is always going to end up costing the organisation more money.
Just use of peripherals alone – and for business use that’s mainly scanners, printers, security tokens and thumb drives – can cause major disruption for a desktop virtualisation project if not handled correctly from the outset.
From the outside in, not the inside out
The primary benefits of desktop virtualisation – as with server virtualisation – comes from the centralisation of IT. However, while server virtualisation should be planned and deployed from the centre of the IT system (primarily data centres), the approach for desktop virtualisation has to be from the outside in.
That "outside" begins with individual employees, their computing needs and the devices they use. The mindset should be that the primary sign-off for desktop virtualisation has to come from the end-users.
Once viewed from the outside in, the strategy, planning and implementation of a desktop virtualisation initiative becomes far more logical. By understanding what is needed at the end-user level, the benefits of centralised management are quickly realised.