AtNet: A Novel Anti-tracking Network with Multi-Party Judgement Capability Based on Cross-Domain Small-World Topology

Anti-tracking network plays an important role to protect the participants from being associated in a conversation. Centralized networks are probably blocked or leak users’ communication privacy. So anti-tracking network is usually a decentralized P2P overlay network. It needs good network robustness. The existing anti-tracking networks are lack of defense capability. A malicious node can easily find the key points or even all participants in the network systems. To address these problems, we propose AtNet, a novel anti-tracking network based on cross-domain small-world topology. First, we generate a small-world topology with more uniform domain distribution based on greedy thought. Then we show a recursive requesting method to maintain the network’s properties. Last, we explain the multi-party judgement mechanism. With the collaboration of near nodes, a node’s abnormal requests can be found. Compared with three state-of-the-art anti-tracking network topologies, AtNet obviously has better robustness. Moreover, we implement the prototype system and evaluate the maintenance effect and defense effect. The experiment results show that AtNet can keep good robustness, anti-tracking capability, small-world property and high clustering when nodes randomly join in or exit the network, and a node can only averagely detect 7.71 nodes in the network with 1000 nodes when allowing nodes to lie twice.