Perpetual Codes for Network Coding

Abstract: Random Linear Network Coding (RLNC) provides a theoretically efficient method for coding. Some of its practical drawbacks are the complexity of decoding and the overhead due to the coding vectors. For computationally weak and battery-driven platforms, these challenges are particular important. In this work, we consider the coding variant Perpetual codes which are sparse, non-uniform and the coding vectors have a compact representation. The sparsity allows for fast encoding and decoding, and the non-uniform protection of symbols enables recoding where the produced symbols are indistinguishable from those encoded at the source. The presented results show that the approach can provide a coding overhead arbitrarily close to that of RLNC, but at reduced computational load. The achieved gain over RLNC grows with the generation size, and both encoding and decoding throughput is approximately one order of magnitude higher compared to RLNC at a generation size of 2048. Additionally, the approach allows for easy adjustment between coding throughput and code overhead, which makes it suitable for a broad range of platforms and applications.