Performance Analysis of Large Scale Peer-to-Peer Overlays using Markov Chains

Emmanuelle Anceaume (1) , Bruno Sericola (2) , Romaric Ludinard (3) , Frédéric Tronel (3)

1 ADEPT - Algorithms for Dynamic Dependable Systems
2 DIONYSOS - Dependability Interoperability and perfOrmance aNalYsiS Of networkS
3 SUPELEC-Campus Rennes

In this paper we consider the problem of targeted attacks in large scale peer-to-peer overlays. Targeted attacks aimed at exhausting key resources of targeted hosts to diminish the target node capacity to provide or receive services. To defend the system against such attacks, we rely on clustering and implement induced churn to preserve randomness of nodes identiers so that adversarial predictions are impossible. We propose robust join, leave, merge and split operations to discourage brute force denial of services and pollution attacks. We show that combining a small amount of randomization in the operations, and adequately tuning the sojourn time of peers in the same region of the overlay allows to decrease the eect of targeted attacks at cluster level, but more importantly to prevent pollution propagation in the whole overlay.

Mots clés: Markov chains; Clusterized P2P Overlays; Adversary; Churn; Collusion; Markov chains.

Domaines: Autre [cs.OH]

Emmanuelle Anceaume, Bruno Sericola, Romaric Ludinard, Frédéric Tronel. Performance Analysis of Large Scale Peer-to-Peer Overlays using Markov Chains. [Research Report] PI-1963, 2010, pp.18. ⟨inria-00546039⟩