On the performance and fairness of BitTorrent-like data swarming systems with NAT devices

There is no doubt that BitTorrent nowadays is one of the most popular peer-to-peer (P2P) applications on the Internet. On the other hand, Network Address Translation (NAT) has become pervasive in almost all networking scenarios. Despite the effort of NAT traversal, it is still very likely that applications, especially P2P ones, cannot receive incoming connection requests properly if they are behind NAT. In this paper, we build analytical models to characterize BitTorrent-like P2P systems with the presence of homogeneous and heterogeneous NAT peers, from which we observe and mathematically reason the poor download performance of NAT peers. Motivated by this observation, we further investigate the fairness metrics of the system, which reveal the fundamental problem of the BitTorrent unchoke mechanism: it is unaware of, and unfair to NAT peers. We thus propose a tunable optimistic unchoke strategy, in order to improve the overall system performance and fairness metrics considerably. The simulation results have validated our analytical models, and also the efficacy of our tunable optimistic unchoke strategy.