Minimum Latency Gossiping in Wireless Sensor Networks

Gossiping is one of the most crucial applications in Wireless Sensor Networks (WSNs) which has been the focus of many researchers. A main issue of gossiping is how to assign timeslots to nodes for interference-free data transmission. There are three models concerning gossiping in WSNs: unit-sized, bounded-sized, or unbounded-sized messages. For these models, the problem of constructing minimum latency gossiping schedules has been widely studied in the literature although most of the existing studies are based on the graph model. In this paper, we study the Minimum Latency Gossiping (MLG) problem with unbounded-sized messages in the graph model as well as the more realistic physical interference model known as Signal-to-Interference-Noise-Ratio (SINR) where there exist relatively few works [1] on the three gossiping models. In the SINR model, we prove the NP-hardness of the MLG problem with unbounded-sized messages. In both the graph model and SINR model, we propose a constant factor approximation algorithm that yields schedules whose latency is bounded by O(∆ + R), where ∆ is the maximum node degree of a network and R is its radius. We also study the performance of the algorithm through simulation.