Data management is critical to the functioning of wide-area wireless communication networks. The communication infrastructure maintains user profiles to keep track of information for operations such as call setup and delivery, subscriber authentication, and billing. Data management deals with the problem of efficiently managing the information in user profiles, which may be distributed across the network. In evaluating different data management schemes, performance measures of interest are: 1) profile lookup and update response times; 2) individual and global database access requirements: 3) individual communication link and global signaling bandwidth requirements; 4) total memory requirement; and 5)overall system cost. Information in user profiles can be classified into "system information", such as current locations of subscribers and subscriber authorization information, "service provider information", such as billing records, and "feature support information", such as numbers for call screening or call blocking. These different categories of information have different demands on their data management schemes. One of the primary objectives of our proposed research is to investigate and develop novel database architectures and protocols for efficiently managing all the classes of information in user profiles. Specific schemes will be designed to meet the various data management requirements. We are especially interested in developing schemes that, unlike current schemes, provide lifelong user numbers through the support of location independent numbering. Scalability and backward compatibility to existing standards, IS-41 and GSM, are also important design issues. We have conceived a promising architecture that is based on a hierarchical organization of databases. Its performance will be compared with a representative set of other data management schemes through extensive simulation studies using a realistic software simulator that we are developing. Another major objective of our proposal is to develop a realistic and extensible simulation platform for studying the performance of data management schemes. Realistic user call and movement models are crucial to our performance evaluations. We will develop detailed call and movement models and corroborate them using actual traffic data. We have laid the foundation in this area through the development of the first version of a highly configurable software simulator, Pleiades. Our research in user behavior models for our simulations is expected to have applications in other areas of wireless communication research.
