Multi-shot network coding

Abstract

The seminal paper of Koetter and Kschischang [11] introduced coding concepts for errors and erasures in a random network coding setting and since then has opened a major research area in communication technology. Here, the network is allowed to change very quickly, which is the case in many mobile applications. The problem is suitably modeled via the operator channel, which makes a very clear connection between network coding and classical information theory. However, coding can also be performed over multiple uses of the network, whose internal structure may change at each shot, giving rise to the so-called multi-shot network coding. Although the potential of using multi-shot network coding was already observed in [11], only recently this more involved approach have been investigated. The general idea stems from the fact that creating dependencies among the transmitted codewords of different shots can improve the error-correction capabilities. A very natural way to impose correlation of codewords(subspaces) over time is by means of convolutional codes, originating the notion of rank metric convolutional codes. In this Chapter we review some of the main results and ideas of multi-shot network coding, propose new approaches and point out some avenues for future research.