Detecting and mitigating P2P eclipse attacks

Peer-to-Peer (P2P) protocols increasingly constitute the foundations for many large-scale applications as the inherently distributed nature of P2P easily supports both scalability and fault-tolerance. However, the decentralized design of P2P also exposes it to a variety of distributed threats with Eclipse Attacks (EAs) being a prominent type to impact P2P functionality. While the basic technique of divergent lookups has been demonstrated for suitability to mitigate EA, it can only (effectively) address limited variants of EAs. This paper investigates both the detection and mitigation potential of enhanced divergent lookups for handling complex EA scenarios. In addition, we propose an approach that can identify malicious peers with a high degree of accuracy. Our simulations have shown EA mitigation rates of up to 96% in case 25% of the peers are malicious. Also, our approach allows for anonymity-fostering, fully decentralized usage, and facilitating downstream mechanisms such as malicious peer removal. © 2015 IEEE