Maximum rank distance profile codes

Abstract

Network coding theory provides a pragmatic instrument to disseminate information (packets) over networks where there may be many information sources and possibly many receivers. From a mathematical point of view, these packets can be modelled by columns of matrices over a finite field Fq and during the transmission, these columns are linearly combined at each node of the network. To achieve reliable communication over this channel, rank-metric codes are typically employed. Nevertheless, network coding techniques for streaming are fundamentally different from the classical ones. To be optimised they must operate under low-latency, sequential encoding and decoding constraints, and as such they must inherently have a convolutional structure. That is the reason why most of the proposed schemes for this scenario employ convolutional codes in different ways.