Location and internal nodes based routing algorithms in wireless networks

Several distributed routing algorithms for wireless networks were described recently, based on location information of nodes available via Global Positioning System (GPS). In GEDIR algorithm (shown to have close performance to the shortest path algorithm, if successful), sender or each node currently holding the message m forwards m to one of its neighbors which is closest to destination. FACE algorithm guarantees the delivery of m if the network, modeled by unit graph, is connected. GFG algorithm combines GEDIR and FACE algorithm. GEDIR algorithm is applied as long as possible, until delivery or a failure. In case of failure, the algorithm switches to FACE algorithm until a node closer to destination than last failure node is found, at which point GEDIR is applied again. In this paper we further improve the performance of GFG algorithm, by reducing its average hop count. Each node in the network is classified as internal or not, based on geographic position of its neighboring nodes. The network of internal nodes define a dominating set, i.e. it must be connected, and each node must be either internal or directly connected to an internal node. We apply several possible definitions of internal nodes, namely the concept of intermediate and gateway nodes from literature, and propose to apply shortest path length between two nodes at distance 2 as additional criterion for internal nodes. A node A is an intermediate node if there exist two neighboring nodes B and C such that A is on a shortest path (in terms of hop count) between them (i.e. these two neighbors are not directly connected). The number of intermediate nodes can be reduced by introducing a secondary measure of a path length. Routing is performed on the dominating set, except possibly the first and last hops. We propose to run GEDIR algorithm on the dominating set, resorting to FACE algorithm whenever GEDIR fails. The performance of proposed algorithms that restricts GFG to internal nodes is also measured in this paper, by comparing its average hop count with hop count of the benchmark shortest path algorithm.