Network Coding for Overhead Reduction in Delay Tolerant Networks

Abstract

Employing Network Coding (NC) in routing protocols for Delay Tolerant Networks (DTNs) is a challenging issue during the last years due to its potential beneficial effect on improving relevant critical metrics. This paper proposes an effective NC scheme intended for sparse DTNs comprising nodes of limited storage capacity. The scheme employs a memory management algorithm that makes optimal use of the limited storage capacity and focuses on unicast sessions where source and intermediate nodes combine only packets belonging to the same generation and destined for the same destination node. Upon having received a sufficient number of linearly independent packets resulting in a matrix of full rank, the destination nodes can apply network decoding to retrieve the information intended for them. The proposed scheme is fully distributed since the network nodes make decisions based on information collected solely from their own buffers. Combined either with Epidemic or probabilistic routing protocols, the scheme manages to drastically reduce the overhead created per source packet.