Network Coding in Disruption Tolerant Networks

In recent years, wireless communication technologies have been increasingly deployed in challenging environments where there is no communication infrastructure, as evidenced by the many efforts in building and deploying wireless sensor networks for wildlife tracking [22, 43], underwater sensor networks [38, 41], disaster relief team networks, networks for remote areas or for rural areas in developing countries [8, 10, 51], vehicular networks [4, 20] and Pocket Switched Networks [19]. Without infrastructure support, such networks solely rely on peer-to-peer connectivity among wireless radios to support data communication. Owing to limits of transmission power, fast node mobility, sparse node density and frequent equipment failures, many of such networks have only intermittent connectivity, and experience frequent disconnection of nodes. Disruption Tolerant Network (DTN, or Delay Tolerant Network), refers to such a network where there is often no contemporaneous path from the source node to the destination node. End-to-end communication in DTNs adopts a so-called “store-carry-forward” paradigm a node receiving a packet buffers and carries the packet as it moves, passing the packet on to new nodes that it encounters. When the destination node meets a node that carries the packet, the packet is delivered to the destination. In addition to intermittent connectivity and dynamic topologies, routing in DTNs faces additional challenges due to the severe resource constraints: for the small mobile nodes carried by animals or humans, buffer space, transmission bandwidth and power are very limited; for mobile nodes in vehicle based networks, even though buffer space or power are usually not severely constrained, transmission bandwidth is still a scarce resource. There has been a substantial amount of research on the benefits of network coding for wireless networks. For multicast applications in static wired or wireless network, Lun et al. [35] and Wu et al. [52] showed that for the problem of minimum-energy multicast, the use of network coding simplifies the problem (from an NP-complete problem to a linear optimization problem solvable