 |
The hypothetical debate between the CFO and CIO highlights a major issue: operational efficiency is all about achieving a fine balance between demand and supply of IT resources, while being able to quickly react to the short-term peaks suggests the need for some idle capacity. So who is right? Actually, both of them are (but, as we’re an IT-focused magazine, let’s say that the CIO is probably more right).
The CFO is already under pressure to better utilise assets – the last thing he wants is more computing resources dragging down his percentage metrics and ruining the overall enterprise ROI. On the other hand, the CIO is more practical. He recognises that there is no such thing as a perfect plan; at times of greater demand from business and users, IT has to be able to respond or the business suffers.
Many observers have begun to question the traditional mindset of IT ownership and taken inspiration from the way we buy utilities such as water, electricity and gas. Computing power on-demand, like turning on a tap or flicking a switch, is a beautifully simple concept.
Enterprises connect to an external resource when they need it, pay for that use, then disconnect when the task is finished. The concept is known as utility computing, and it is changing IT from a capital investment that you own, operate and manage in-house, to a service that you access externally, as and when you need it.
So how does the model work in practice? Are businesses shutting down their data centres and outsourcing all their IT needs? No, the utility computing model is more of a complement to what businesses already have. It’s designed to give customers access to additional IT resources, quickly and cost-effectively, on a temporary basis – so they can smooth peaks in demand.
Traditionally, enterprises that need to drive one-off development projects, have had to purchase new servers in order to meet those short-term peaks. This is a lengthy process that involves calculating the exact requirements, negotiating the best deal, ordering and taking delivery, unpacking, installing, configuring and integrating the equipment.
The utility model allows them to connect straight into an existing data centre, run by a third party, where all the resources they need are ready to run. The project can be sprung into life almost instantly. Then, when it is complete, they can simply “turn off” the supply. They only pay for the computing time used, and they are not left with a mass of costly under-utilised hardware. It also means they don’t have to invest in extra data centre space, which is costly and often hard to source in expensive metropolitan areas.
|
Printable version
