HTL: A locality bounded flat hash location service

Abstract

Many location services have been proposed, but some challenges remain. In this paper, we present a new location service, named HTL (Hash Table Localized) to solve the locality problem, that in a location service, the location information can be stored potentially far away from both the source and destination nodes, even when the source and destination nodes are close. As a result, it causes high overhead in update and query. HTL uses a double index hash function to map a node to a location in the network area called the virtual coordination of that node. Then, a novel method is employed to divide the physical space into lattices. The publish and query algorithms are designed based on this division. In HTL, when the distance between the source and destination nodes is l, the cost of query is O(l 2). We define this property as n 2-locality bounded. HTL is the location service that achieves this property with the least storage and network overhead. Both analysis and experiment results are presented in this paper concerned with the cost, the locality bounded property and the scalability. © 2008 Springer Berlin Heidelberg.