On detecting malicious behaviors in interactive networks: Algorithms and analysis

Abstract

Interactive networks are vulnerable to various attacks due to the existence of malicious nodes which do not comply with the network protocol so as to achieve their own purposes. For example, in a peer-to-peer (P2P) streaming system, since each peer needs to participate in uploading data to other peers, malicious peers may choose to upload bogus data so as to damage the playback and degrade the watching experience of normal peers in the system. This is known as pollution attack in P2P networks, and it can cause severe impact to the performance of P2P streaming systems. Other examples include pollution attack in wireless mesh networks (WMNs) where malicious nodes forward modified and polluted packets to other nodes, and the shill attack in online social networks (OSNs) where malicious users give wrong recommendations to others so as to mislead their purchases. In this paper, we propose a general and fully distributed detection framework which can be executed by each legitimate node in an interactive network to identify its malicious neighbors. To illustrate the effectiveness and the efficiency of our detection framework, we apply it to three realistic applications: P2P streaming networks, WMNs and OSNs, and show how to defend against the attacks launched by malicious nodes. We also quantify the performance of our detection algorithms and validate our analysis via extensive simulations.