The market for desktop virtualisation is set to grow exponentially over the next few years. Gartner have produced some much quoted figures that claim the market will be worth $65 billion in 2013 – which is all very good, but does not necessarily help a hard-pressed IT leader make the case internally for an investment in technology that has been around for some time, but is still considered niche by most.
Part of the problem is the perceived ‘cheapness’ of PC’s, vs the relatively expensive nature of a most VDI solutions out there. The price of PC hardware has fallen dramatically in recent years, as any walk around your local PC World will tell you. For £500 to £600 it is possible to buy an enormously capable PC, complete with an OS. For corporate customers who usually specify their desktop machines quite highly, in an effort to make them last at least three years with a degree of future proofing, perhaps the cost per PC might be as high as £1000 to £1200, which translates to a monthly cost of £28 to £33 per user per month over 3 years.
This means that on a hardware comparison alone it is hard to justify the £ millions an on-premises VDI project or even the subscription costs of a desktops-as-a-service (DaaS) solution from someone like Molten Technologies, where you can buy a ‘virtual container’ (the equivalent compute, memory and storage of a PC, but in the data-centre) for circa £30 per user per month.
The problem with this analysis is that it completely ignores the actual costs of running a desktop estate – hardware and software maintenance and support, hardware fixing and the myriad of end-user costs that corporations deal with under the banner ‘user support’. These costs are usually exasperated by poor or even no management of the desktop, for instance having unlocked desktops where users can load their own software and applications and, consequently, have licence to break things (and, of course, then expect the company to pick up tab to fix it for them). Laptops are notoriously fragile and are prone to be dropped, left on trains, have coffee poured on them and so forth, all of which creates downtime when users are either, in the worst case, unable to work at all or are less productive because they lack their normal resources.
This non-productive time is probably the most costly part of any IT failure, yet is rarely measured or quantified. A recent article by Dell put the cost of this lost productivity at $294 per year per user – suddenly a desktop environment that all but removes that cost can look very attractive. Add to that claimed savings of $250 per year for reduced calls to the service desk and desk-side support and the case is genuinely compelling. The issue is, of course, that these savings are relative to the current costs of an organisation and those are generally very poorly understood. Costs reside in different budgets, sometimes in different parts of the organisation, so it is often hard for the IT evangelists to be heard, or make the case when they are.
It is not always easy for the desktop team to claim savings in other budgets. A good example is energy savings – the IT department are rarely the ‘owners’ of the energy bill – so measuring and ‘banking’ those savings can be a challenge. Those savings are real however – thin-client computing devices typically use 10% to 15% of the power of a PC or desktop, as they have relatively tiny amounts of processing capacity and no local hard-drive or fan. In a virtual environment the data-centre does of course use power and cooling, but this is generally in a controlled and optimised environment. Even including DC power use, the total consumption of a thin-client + DC estate vs a physical laptop/desktops is over 75% less. This translates to a total saving of circa £40 per annum per PC, a material sum in an organisation with many thousands of devices.
The organisations that understand their end-user computing costs best are typically those that have outsourced their desktop services and support to a third-party, who then present their clients with all their costs in a single invoice, with neat line items over which they have spent many happy hours negotiating. Perhaps there is a lesson in this – if all organisations treated their desktop provision in the same way they treat an outsourcing exercise, then they would have to consider all the associated costs. This can be summarised in a ‘total cost of ownership’ analysis, beloved of cloud computing consultants – and they are right.
The broader point here is simply not to treat desktop virtualisation as a hardware replacement exercise, but rather as a new way of working that is significantly more secure, makes life easier and more productive for the user …. and just happens to save the company money as it does so.