Analysis and optimization of routing trust values in a Kademlia-based Distributed Hash Table in a malicious environment

Trust management in peer-to-peer (P2P) networks is a current topic. Most of the current work focuses on a “clean routing table”, aiming to keep it free from malicious nodes. We follow a different approach which can be characterized as “clean routing process”: All kinds of nodes may be contained in a node's routing table. During the routing process, only trustworthy nodes shall be used for determining the next hop and finally for storing or retrieving content items. The trust values are stored in a distributed way and the ratings must be given and stored verifiably. In this paper, we present the results of implementing our trust concept into a Kademliabased Distributed Hash Table (DHT). We analyze the distribution of the routing trust values and introduce malicious nodes into the network. The results show that our concept is able to identify them precisely.