A Social-Enhanced Data Verification Framework against Pollution Attacks in P2P Live Streaming

With successful deployments of commercial swarm-based P2P streaming systems, peer-to-peer streaming traffic grows rapidly in the last few years. To leverage system-wide bandwidth, peers in P2P streaming systems are expected to exchange their own data with other peers, and this is extremely vulnerable to pollution attacks, where malicious peers inject fake data blocks into the network and those fake data blocks will soon be propagated by the innocent nodes received them. Such attack leads to degradation of data delivery ratio and failure to decode the streaming data. In this paper, we propose a novel Social-Enhanced hash-based Cross-verification framework (SEC) to defense content pollution attacks. That is, the fake data is able to be rapidly identified through the SEC scheme, and thus the malicious nodes are isolated from clean peers. With the assistance of social networks, we can reinforce the resilience of our SEC scheme for multiple colluding malicious peers and improve the data availability by the AIMD-based reputation module. The results show that the SEC framework is able to provide high delivery ratio, (as well as high accuracy, low computational overhead, low communication overhead, and low delay) for fulfilling the stringent requirements of P2P live streaming systems even under a severe or dynamic environment (e.g., when the total capacity of polluters is larger than the source node capacity).