Topology-Aware Hybrid Random Walk Protocols for Wireless Multihop Networks

Abstract

The proliferation of wireless multihop networks has made various operations, such as search and retrieval of distributed data a significant concern. Various methods have been proposed for performing such tasks efficiently, especially when all network nodes need to be visited at least once. Random walks are probabilistic approaches for performing the aforementioned operations effectively and with relatively small overhead compared to other typically-employed schemes, such as flooding. Recently, a hybrid random walk scheme has been proposed for increasing the desired performance, at the cost of additional consumed resources. In this work, we adopt the paradigm of hybrid random walk protocols and propose two novel hybrid schemes that exploit local topological information, aiming at further increasing the performance of random walk protocols in multihop networks. We consider different jump configurations of the hybrid random walk protocols and various degrees of mobility. Through analysis and simulation, the simple random walk model appears more appropriate for energy-constrained networks such as sensor networks, while the hybrid ones are more appealing for less energy-stringent, performance-oriented multihop networks, such as vehicular and mesh networks. The simple hybrid protocol occupies the middle ground, being appealing for ad hoc networks with medium to low node densities and average energy requirements.