When you put on a few kilos you let out your belt to accommodate. When your business booms, your computer systems should adapt as easily. Helen Bradley explains how to establish a system that will grow or shrink as required in the future.

Any I.T. vendor will tell you, one of the key issues to consider in an IT purchase is scalability. In other words, room to grow or change in all directions, including shrinking. It sounds simple, but too often businesses get trapped in the situation where the limits of software or hardware are reached well before the end of the system's lifecycle. So, how can you avoid this?

Scalability means different things to different businesses. In some cases scalability is related to the changing needs and changing workforce in the organisation. In others, it involves growing the scope of the services you use-for example, starting out with a general accounting system and later adding inventory or payroll. Scalability is also relevant to hardware. It's all about having room to grow in your computer system without having to replace it because you've outgrown it or because it can't be extended.

When there's more work to be done and more people are needed to do it, the computer system that supports this work must be able to be scaled upwards in a timely manner to accommodate the changes. In some cases this growth will ebb and flow following a seasonal cycle, so the system must have the ability not only to grow but also to be scaled back as needed. In other situations the growth will be all positive and the systems will need to be able to continue to cater for this growth, whether it occurs within your current business or involves expanding into new geographical locations.

For many businesses, the scalability needs of the organisation means growing what the system does. You may start with a system that performs one discrete task and later extend it so it performs a wider variety of tasks. When the system grows by adding applications, this change must be managed intelligently so, for example, the existing data can be used and the new system can be integrated with existing systems. It's not advisable to have each new system spawning its own duplicate data store-the costs of having to keep them updated and the risks associated with them not being updated are high.