Advances in Iterative Decoding and Maximum Likelihood Decoding for the Packet Network with Comparisons to Fountain Codes over the Erasure Channel

In this paper, we propose a novel ML decoding algorithm - the In-place Algorithm in conjunction with a Product Packetisation method for the congested Internet Network modeled as a Binary Erasure Channel (BEC). Any code can be used with this algorithm and we give results for cyclic codes constructed from BCH codes. Existing codes and decoding algorithms are compared in terms of performance and decoding complexity. It is shown that a significant performance improvement can be achieved. In general, ML decoding on the Erasure channel has a complexity of O(N3) or O(N2) depending on the algorithm. It is shown that this complexity can be reduced to O(N1:5) for the network channel by using the product packetisation method. With an analysis in performance, ratelessness can be achieved by acks.