Self-adapting Linear Network Coding Emulation

Abstract

Linear network coding (LNC) introduced a new paradigm for routing data across networks where the transported packets are not the original information packets but rather linear combinations of the original packets. This is known to allow a more efficient use of the network resources. With LNC based on systematic network codes, the network's capacity is chiefly defined by the rank of the end-to-end equivalent transfer matrix. This paper presents a network-layer emulator based on the obj ect-oriented programming paradigm, which allows, for any network topology, to have the network's capacity self-optimized by adapting the coding operations performed at each node, depending on the impact the changes have on the global rank of the global transfer matrix, and on a metric related to the use of network's resources.