Browse by author
Lookup NU author(s): David Cooper
Full text is not currently available for this publication.
Mobile Ad-hoc Networks (MANETs) allow deployment of mobile wireless devices or nodes in a range of environments without any fixed infrastructure and hence at a minimal setup cost. Broadcast support that assures a high coverage (i.e., a large fraction of nodes receiving a broadcast) is essential for hosting user applications, and is also non-trivial to achieve due to the nature of devices and mobility. We propose Encounter Gossip, a novel broadcast protocol, which holds minimal state and is unaware of network topology. Coverage obtained can be made arbitrarily close to 1 at a moderate cost of extra message traffic, even in partition-prone networks. Under certain simplifying assumptions, it is shown that a high coverage is achieved by making a total of O(n ln n) broadcasts, where n is the number of nodes, and the time to propagate a message is O(ln n). The effect of various network parameters on the protocol performance is examined. We then propose modifications to minimise the number of redundant transmissions without compromising the achieved coverage. Two approaches are pursued: timer based and history based. The effectiveness of each of these approaches is assessed through an extensive set of simulation experiments in the context of two mobility models. Specifically, we introduce a new heuristic alpha policy which achieves significant reduction in redundancy with negligible reduction in coverage. A generalisation to multiple broadcasts proceeding in parallel is proposed and the protocol is refined to reduce problems that can occur due to the effects of high mobility when transmitting a large number of messages. Finally, we implement and validate Encounter Gossip in the context of a real-life mobile ad-hoc network. All these investigations suggest that the protocol, together with the proposed modifications and refinements, is suited to MANETs of varying degrees of node densities and speeds.
Author(s): Cooper DE
Publication type: Report
Publication status: Published
Series Title:
Year: 2008
Institution: School of Computing Science, University of Newcastle upon Tyne
Place Published: Newcastle upon Tyne
