A Packet Encoding Algorithm for Network Coding with Multiple Next Hop Neighbor Consideration and its Integration with Delta Routing

Abstract

We propose an efficient packet encoding algorithm for network coding that considers multiple neighbors, as the next hop candidates for a packet while searching for an optimal packet combination to transmit. Our algorithm, based on bipartite graphs, lets a node exhaustively search its queue to identify the maximum set of packets that can be combined in a single transmission. This algorithm can be easily integrated with a routing scheme in which a node considers multiple neighbors before choosing to forward a packet to one of them. We demonstrate this by coupling our algorithm with delta routing. Throughput benefit offered due to network coding is often hampered by congested nodes in a network. Integration of our encoding algorithm with delta routing helps packets circumnavigate congested parts of the network, as demonstrated by the contour graphs of transmission densities. In our routing mechanism a node only requires information available from its immediate neighbors in order to make decisions regarding coding and forwarding the packets. Our simulations show that by encoding packets more efficiently and routing them through the underutilized parts of the network, our routing scheme achieves 50.00% to 92.59% higher throughput than conventional shortest path routing. This throughput benefit is 23.68% to 55.88% improvement over utilizing coding on the shortest path routing.