Reorganization of Clusters of Similar Operations

In order to ensure the high availability of web services operations; recently we have opted for a communities-based approach. This approach consists in grouping in a single cluster all operations offering the same functionality (the same outputs). The client (for a substitution requirement e.g.) can select, in this set, the operation (s) according to their non-functional criteria but also according to their inputs parameters. The customer is satisfied, on the one hand, he has all similar operations as the failed one, but, on the other hand, he feels lost: what are the operations taking this or that inputs, as the failed one? In this context, and in order to guide and help users in their choices, we have considered the reorganization of the clusters by specifying the relations linking the inputs of the operations of the same cluster.