Performance of Semantic-dependent Two-tier Gossip Mechanisms

The efficiency of locating heat sources in a sensor network with the novel semantic-dependent two-tier gossip mechanisms is studied. Two-tier gossip mechanisms are probabilistic paradigms which can switch from one gossipee-selection scheme to another one, typically one that tends to select more nearby nodes. An upper time bound for two-tier gossip mechanism is presented. When semantic-dependent two-tier gossip mechanisms are employed, the selection of communication peers depends on information, for example temperature, obtained from the application layer. Distance from the heat source(s) can be estimated with the help of the local temperature detected, and the two-tier technique is also employed for further improvement. With uniform gossip as the benchmark, an overall improvement in both time and the communication distances required is observed.