On-demand / Utility Computing
On-demand computing is a concept which seems to fit well with the concept of a lean, agile enterprise. Essentially it refers to a model whereby a company flexibly adjusts the computing power and capacity it needs through an agreement with a third party that acts as a 'utility', releasing or deploying computing power as it is needed.
Obvious beneficiaries of such an approach are organisations that have a great degree of seasonality in their business or who have one-time annual peaks in capacity demand. Customers must realise, however, that in practice there are trade-offs - for example they lose their ability to design, build and deploy the most appropriate infrastructures. They must take the equipment provided by their supplier - and the configuration, the firewall, etc.
The concept is only in its infancy and, in truth, many of the components are not really in place. For example, the only way in which a supplier can make this profitable is to share equipment amongst customers, lead-balancing as the various peaks and troughs occur. The hardware and software to manage this seamlessly and efficiently is not yet available. Thus, suppliers are operating a currently inefficient model in the hope that the eventual size of the market, and developing software tools, will enable real profits in the future. It has attracted the big players - such as IBM and SUN - who can play this long game.
The real 'holy grail' is to provide the computing power for virtual organisations - small, flexible, low infrastructure companies that may be short-lived in their current manifestation.
January 2004
