Clusters consist of two kinds of computers: servers and any number of workstations (workers). The servers provide the necessary software and the workers receive from the server the data to be processed. The servers controls which worker is currently loaded with work and which one is not. This allows you to process large amounts of data, especially small units of data, much faster.

This kind of processing gives you another advantage but sheer improved processing speed: the system stability of the whole system also improves to a large extent. Until now the processing of data has been carried out on a single computer, if that computer failed the whole processing would stop. By using a greater number of computers, that are used only for the data processing, you reduce this problem immensely because if one worker fails you only have to divide the workload to the other workers. The processing in itself is not stopped, though.