Improving the BitTorrent protocol using different incentive techniques.

摘要

Peer-to-peer (P2P) content sharing protocols dominate the traffic on the Internet [IPO09], and thus have become an important piece in building scalable Internet application. Peers in P2P network are typically independent entities that together form a self-organizing, self-maintaining network with no central authority. As a result, P2P network performance is highly dependent on the amount of voluntary resources individual peers contribute to the system. However, peers in P2P systems have self-interest to control their degree of collaboration and contribution [SP03, FS02, Pap01], as cooperation may incur significant communication and computation costs. Thus, rational peers may refuse to contribute their fair share of resources [AH00, SPD02, HCW05]. In some scenarios this may lead to the "tragedy of commons" [G68], when maximizing peers' own utilities may effectively decrease the overall utility of the system. Hence, mechanisms that incentivize peers to actively cooperate and contribute their resources are fundamental and crucial for these systems' continued success. In such mechanisms, fairness among the peers participating in content distribution is an important factor, as it incentivizes peers to contribute their resources.;Recent research efforts have shown that the popular BitTorrent P2P protocol [Coh03) does not strictly enforce fairness. Moreover BitTorrent allows free-riding, in spite its tit-for-tat mechanism. In an effort to address the incentive-related problems in the BitTorrent protocol, we propose and examine three different protocols that encourage cooperation and contribution of resources, and improve fairness and resistance to free-riding in BitTorrent-like systems.;We propose the team-enhanced Bit Torrent protocol that dynamically organizes peers of similar upload bandwidth in teams---groups of peers that are formed by a central entity. Based on an analytical model, we prove that the dominant strategy of a rational peer in a system with teams is to be a team member. In our experiments, we observed download time improvement of 10%--26% for all team members, while free-riders were delayed by more than 100%.;Additionally, we propose the buddy-enhanced Bit Torrent protocol , which distributively and dynamically creates buddies---pairs of peers having similar upload bandwidth whom collaborate for mutual benefit. We prove the existence of Nash Equilibrium when all of the contributing peers adopt the buddy protocol. In our experiments, we observed improving fairness in the buddy-enhanced network where high capacity peers improved their download time by 8%--36%. Moreover, in a network with free-riders, all buddy-enhanced leechers improved their download time by 2%--36%, while free-riders were delayed by 21%--48%.;Finally, we propose the foresighted resource reciprocation protocol . Here we use a reinforcement-learning process to capture the associated peers' statistical behaviors and their corresponding expected utilities. In our experiments, the protocol improves fairness, where high-capacity peers improved their download time by up to 33%, while delaying free-riders by 8%--20%.