High throughput database structures for location management in PCS networks

This paper introduces a distributed hierarchical database architecture for location management in global personal communications system (PCS) networks with the location-independent numbering plan. The proposed structure consists of a number of database subsystems, each of which is a three-level tree structure and is connected to the others only through its root. Due to the localized nature of calling and mobility patterns, this scheme effectively reduces the database loads and signaling traffic incurred by the location update and call delivery procedures. Moreover, based on the flat numbering plan, two memory-resident access structures-the memory-resident direct file and the T-tree are proposed for the location databases to further improve the database throughput. Analytical models and numerical examples are presented to evaluate the response time and the storage capacity required for each location database. Results show that our method can produce a high-throughput database system, which is crucial for the deployment of high-capacity PCS systems in the future. In addition, modifications to the data structure, the insert algorithm, and the delete algorithm for the T-tree are proposed to enhance the performance of the T-tree.