Comparison of FEC types to optimize TCP connection over AWGN satellite channels

The optimisation of the end-to-end throughput of a TCP connection over geostationary satellite links is a challenging research topic. This because the high delay-bandwidth product, together with a non negligible random loss of packets, are conditions which differ pretty much from the original environment TCP was originally designed for. As a result TCP performance is significantly impaired by the channel bit error rate. The literature is full of suggestions to improve TCP goodput, most of them are based on modifications of the protocol itself. We only investigate on the application of different FEC (forward error correction) types for TCP optimisation, leaving the end-to-end protocol unaltered. Using a method, which is a midway between analysis and simulation, to evaluate the throughput of a TCP long run connection, we first disprove the widely diffused opinion that the packet loss rate, introduced by errors on the channel, should be negligible with respect to that caused by congestion. We then compare physical-level FEC techniques such as convolutional encoding/Viterbi decoding and Reed Solomon, link-level erasure codes and their combinations, over a wide field of signal to noise conditions of the satellite channel. Furthermore, the case of multiple connections per link is compared with the single connection per link case.