A hierachical location management architecture for wireless ATM networks

This paper proposes a hierarchical location management scheme for tracking roaming users in an ATM based mobile network. Most of the existing location techniques are based on a near-flat classical scheme which relies on two types of location databases, namely, the home location register (HLR) and visiting location register (VLR). Distributing mobile users' location in formation only to their respective HLRs and VLRs makes these schemes scalable in terms of the network storage requirements. However, the update and the query costs are not quite optimized in these near-flat policies. Especially, in scenarios where every short-distance local move is required to be reported to a distant HLR, the schemes suffer from the update-scalability problem. Similarly, while located near the callee when a caller needs to go to a distant HLR of the callee, the query cost is unnecessarily amplified. To avoid these problems, caused primarily by the flatness of the policy, we introduce a location hierarchy built on top of physical ATM network topology. We show that a hierarchy of location agents, together with the classical HLRs and VLRs, can reduce the update and query costs while keeping the policy storage-scalable. We also show how a caching scheme can be effectively implemented on top of these location hierarchies for further reducing the query and update costs. An exhaustive simulation model, for evaluating the performance of the baseline and the caching scheme, was built. Simulation results show that in the presence of reasonable spatial locality in both user mobility and query profiles, our scheme can offer up to 73% improvement in update performance and up to 27% improvement in query performance over the classical scheme.