Mobile Networks and Applications 6, 117–124, 2001  2001 Kluwer Academic Publishers. Manufactured in The Netherlands. Optimal Distributed Location Management in Mobile Networks ∗ GOVIND KRISHNAMURTHI Department of Electrical and Computer Engineering, Iowa State University, Ames, IA 50011, USA MURAT AZIZO ˜ GLU Department of Electrical Engineering, University of Washington, Seattle, WA 98195, USA ARUN K. SOMANI Department of Electrical and Computer Engineering, Iowa State University, Ames, IA 50011, USA Abstract. An important issue in the design of future Personal Communication Services (PCS) networks is the efficient management of location information. In this paper, we consider a distributed database architecture for location management in which update and query loads of the individual databases are balanced. We obtain lower bounds to the worst-case delay in locating a mobile user, to the average delay, and to the call blocking probability. Wethen propose a dynamic location management algorithm that meets these lower bounds. The optimality of this algorithm with respect to these three performance measures, as well as simplicity, make it an appealing candidate for distributed location management in PCS networks. Keywords: PCS networks, mobile networking, distributed location management, optimal load balancing 1. Introduction Cellular communication technology has advanced rapidly in the last decade with significant developments in the capa- bilities of mobile networks. Future Personal Communica- tions Services (PCS) networks will be capable of providing a broad range of services to a large population. Significant research efforts have being undertaken to address various de- sign problems in PCS networks such as channel allocation, location management, and efficient power management. In this paper we focus on one of these problems, namely, the location management problem. In PCS networks, a location tracking mechanism is needed to determine the position of the mobile hosts 1 in order to establish connections. Current methods require a mobile to report its location to the network when necessary. Some location management schemes use a time-based up- date strategy. Other schemes require a mobile to update its location only if its present location is at least a predeter- mined distance away from its location at the previous up- date. The network stores the location of the mobile in lo- cation information databases (LIDs) and this information is retrieved during call delivery. In a distributed database archi- tecture with multiple LIDs, it is beneficial to balance the load on each database in order to avoid unnecessary performance bottlenecks. Our objective in this paper is to develop and analyze distributed location management algorithms which are not only load-balanced, but also optimal with respect to query delay and call blocking probability. ∗ This research was supported in part by the David C. Nicholas Professor- ship of Electrical and Computer Engineering at Iowa State University. An earlier version of this paper appeared in [20]. 1 We refer to mobile hosts as mobiles in the rest of the paper. The rest of the paper is organized as follows. In sec- tion 2 we describe the network architecture we use in this work. In section 3 we discuss some of the previous ap- proaches to distributed location management. In section 4 we obtain lower bounds to the performance of load-balanced location management algorithms. The construction of a dy- namic load-balanced location management algorithm which achieves these lower bounds is presented in section 5. We conclude the paper in section 6. 2. System model Most PCS networks use a cellular architecture for bandwidth efficiency [1,19]. In this architecture, each cell has a base station to which the mobiles of the cell communicate through a wireless link. A set of base stations is controlled by a Base Station Controller (BSC). The primary function of a BSC is to manage the radio resources of its base stations, by performing handoffs and by allocating radio channels. Each BSC is connected to a Mobile Switching Center (MSC) through a wired network. A MSC typically provides switch- ing functions and coordinates location registration and call delivery. The MSC has access to the location information databases in the network, which are used to store location and service information for each registered mobile of the PCS network. PCS architectures which adopt the IS-41 and GSM standards, [5] and [13], use a two-level hierarchy of such databases for location management as shown in fig- ure 1. These are the Home Location Register (HLR) and the Visitor Location Register (VLR). The HLR is a global database in which information about all registered mobiles. A VLR is a local database usually associated with the MSC