An efficient Soliton-like network coding protocol for the resource-constrained Y-network

Abstract

Originally designed for point-to-point transmissions, fountain codes are capacity-achieving codes over a binary erasure channel. On the other hand, network coding is an optimal multihop data dissemination protocol whose high decoding complexity makes it too expensive for resource-constrained applications. Soliton-like rateless coding (SLRC) has previously combined network and fountain coding paradigms such that a less complex decoder can be applied. Specifically, the coding done at an intermediate relay node allows a belief propagation decoder to be efficiently applied. We extend the SLRC protocol and propose the Improved Soliton-like Rateless Coding (ISLRC) protocol. In ISLRC, the relay applies intelligent coding and shapes the degree distribution such that performance is better than SLRC. In addition, ISLRCs performance gains are achieved using fewer resources than SLRC while maintaining all the key properties of SLRC. Simulation results show that even under the worst-case scenario of ISLRC, better performance can be achieved compared to SLRC and other existing schemes.