Goodput optimization of long-lived TCP connections in a faded satellite channel. Extended Version

The optimization of the end-to-end throughput of a TCP connection over geostationary satellite links is a challenging research topic because the high delay-bandwidth product, together with a non-negligible random loss of packets, are conditions which differ considerably from the original environment for which TCP was originally designed. As a result, TCP performance is significantly impaired by the channel bit error rate. In this paper we investigate the application of different FEC (forward error correction) types/rates and different bit rates, for the optimization of TCP goodput, in transmissions over a rain-faded geostationary satellite channel, provided that the end-to-end protocols are left unaltered. We 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. The case of multiple connections per link is also analyzed, in addition to that of a single connection per link. In order to evaluate the throughput of a TCP long-lived connection we used a method that is midway between analysis (fluid model) and simulation.