METHOD OF CLUSTERING NODES IN A DISTRIBUTED COMPUTER NETWORK

A method of clustering nodes in a distributed computer network to reduce the overall processing and storage requirements at the network nodes. The network is partitioned into subsets (25, 75) with one node (10, 100) in each cluster designated as the cluster control point. The cluster control point represents the cluster as a single node (25 or 75) to the external part of the network. It maintains an internal topology database of network resources within the cluster and an external topology database of network resources outside the cluster for use in determining routes for sessions throughout the distributed network. The cluster control point (10, 100) is the focal point in assisting the external part of the network to locate network resources internal to the cluster and in assisting the internal nodes to locate network resources external to the cluster. Internal resources found as a result of a search are maintained in an internal directory cache at the cluster control point, and external resources found are maintained in an external directory cache.