Skip to Main Content
P2P content distribution has been an important Internet application in recent years. Its popularity draws attentions of Internet Service Providers (ISPs) for traffic management even throttling. This paper presents a new P2P content distribution protocol named ShareStorm that is efficient and ISP friendly. The key is our receiver-driven distance centric heuristics with probabilistic randomization. Compared with BitTorrent, ShareStorm accelerates the content distribution greatly and reduces the operating cost of the ISPs. In addition, it allows the ISPs to adjust the traffic distribution of specified areas dynamically. We conducted both emulation and Internet experiments to demonstrate ShareStorm's performance. Particularly, Share Storm outperforms BitTorrent by more than 60% in our Internet experiments in terms of download time.