Trading Optimality for Scalability in Large-Scale Opportunistic Routing

Opportunistic routing utilizes the broadcast nature of wireless networks, significantly promoting the unicast throughput. Many variations of opportunistic routing designs have been proposed, although all of the current designs consistently rely on all of the topology information to construct forwarder lists and process data forwarding, which indeed restricts the application in large-scale wireless networks, where collecting global optimal information is very costly. In this paper, we propose the localized opportunistic routing (LOR) protocol, which utilizes the distributed minimum transmission selection (MTS-B) algorithm to partition the topology into several nested close-node-sets (CNSs) using local information. LOR can locally realize the optimal opportunistic routing for a large-scale wireless network with low control overhead cost. Since it does not use global topology information, LOR highlights an interesting tradeoff between the global optimality of the used forwarder lists and scalability inferred from the incurred overhead. Extensive simulation results show that LOR dramatically improves performances over extremely opportunistic routing (ExOR) and MAC-independent opportunistic routing protocol (MORE), which are two well-known designs from the literature, in terms of control overhead, end-to-end delay, and throughputs. It also exhibits promising performance in vehicular ad hoc networks (VANETs).