Towards Scalable and Robust Service Discovery in Ubiquitous Computing Environments via Multi-hop Clustering

Large-scale ubiquitous computing environments require scalable and robust service discovery to enable "anytime, anywhere" computing, which is hard to be satisfied in flat network architecture. In this paper, a multi-hop Cluster-based Architecture for Service Discovery (CASD) in ubiquitous computing environments is presented. Based on the Neighborhood Benchmark (NB) which quantifies the connectivity and link stability of mobile nodes, CASD organizes the network consisting of heterogeneous mobile computing devices to be multi-hop clusters, and constructs each cluster to be a local DHT-based p2p network for distributed storage of service indices. The clusterheads are connected together to form a virtual backbone, based on which service discovery messages are disseminated among clusters. It is shown that CASD can control the communication overhead of service discovery when the network scale increases, and that CASD can achieve robust service discovery by maintaining controllable redundancy of service indices in each cluster. The scalability and robustness of our approach in various types of network settings are shown by intensive simulations.