Semi-distributed informed peer selection for scalable P2P overlay network

As an alternative solution to solve the tussles between the network operators and the peer-to-peer (P2P) applications, an informed peer selection (IPS) scheme like P4P is proposed. In IPS, each network operator installs a server that provides their network information to the peers as a guidance for peer selection and each peer chooses its communicating partner based on the guidance. Most existing IPS studies just focus on how they distribute which network information to the peers and they do not focus on how the network information can be provided in a scalable manner. In this paper, we present the design and evaluation of an approach to distribute the network information in a scalable manner. Our approach, called sIPS, utilizes peer neighborship. For this, we define the network information data format including value, scope, and TTL. Through ns2 simulations, we show that sIPS shows better performance than other existing approaches in terms of the network operator server overhead and the traffic overhead.