A cluster is the combination of the index, sequence set and data portions of the dataset. The operating system gives program access to the cluster, ie. to all parts of the dataset simultaneously.