A Cluster Based Service Discovery Model for Mobile Ad Hoc Networks

Abstract

Mobile ad hoc networks, as a typical example of self-organized networks, represent an emerging and promising communication paradigm. Not only the variety of devices but also the diversity of services is continuously increasing. Such services must be provisioned in a flexible and distributed way without a central infrastructure. The discovery of such services is an involved task since it must cope with the high level of mobile device heterogeneity, degree of mobility, and take the limited device resources into account. In this paper, we propose a decentralized and scalable service discovery and invocation framework for ad hoc networks of clustered mobile devices. The basic idea is to use the clusterheads as service directory nodes to assist mobile nodes in the network in finding desired services. In addition to storing the descriptions of services that fall in their clusters, clusterheads also cache descriptions of services that are retrieved from other clusters as they are requested. Experimental analyses are conducted to study the performance of the proposed architecture and to compare it with the broadcast-based model that has been extensively studied in the literature.