An S2P-based resource discovery mechanism for tuple-based pervasive systems

Abstract

A critical issue in pervasive systems is the efficient discovery of resources in dynamic environments. Existing resource discovery mechanisms are inefficient since they broadcast resource request messages to all peers. They are also inflexible and hard to use. By imitating human behavioural patterns in social interactions known as S2P networks we present a more efficient discovery mechanism for tuple-based pervasive systems that only communicates to social peers in search of resources. The choice of tuple-based model yields more flexibility and ease of use. The result of S2P style search for a resource requested by a peer is consistently cached in the local tuple space of that peer for later use. Exemplar scenarios were run on an implementation of the mechanism in Linda in a mobile environment (LIME) on different platforms. The results showed significant improvements in resource discovery and access times.