As telecommunications in the world grow exponentially in importance forbusinesses, the sizes of networks grow, generating increasingly monitored statistics. To handle this data, efficient systems are required to store and manipulate it. This thesis investigates efficient methods to compute, store, and operate the current and historical aggregates, derived from network management data. The networks of interest are those with star architecture.

This thesis proposes a methodology to process the incoming data from the network elements, store it in a database, and then use efficient techniques to perform the aggregation through the various dimensions,namely by attributes, network containment, and time. This procedure generates data with different levels of granularity, which is suitable for the use of modern on-line analysis processing techniques, e.g., drill-downs and scale-ups, and for gradual bottom-up data trimming without losing information, although coarsening it. This thesis demonstrates that modern databases may be tuned and deployed for the purposes of computing and storing network aggregates.