Eliminating Bottlenecks in MANETs

MANETs comprise several mobile nodes, wirelessly connected with each other. Transmission capacities in these networks are limited and easily over-utilized, especially if multiple flows are present. Paths are constructed decentralized, where each participant's routing decision is based on local and incomplete topology knowledge. This makes multi-flow path distribution costly and complex. Nodes of MANETs are mobile and topologies are continuously changing, which on top forces routing to compute paths as fast as possible. Recent work proposes a topology update algorithm, constructing a graph out of the complete network topology at runtime. We leverage this knowledge and propose the pathfinding technique RbDRP, aiming for capacity-conform paths in minimum execution time. Therefore, we artificially generate over-utilized network segments, which have to be resolved by our approach. RbDRP applies a robust link metric to compute long-living paths. Evaluations compare and discuss the results regarding the number of found capacity-conform path settings, the runtimes, and their connection lifetimes.