Measuring the Identifier Repetition and Aliasing Phenomena in K Networks

Kademlia, as a robust distributed Hash table (DHT) protocol, has been deployed by peer-to-peer (P2P) file sharing applications (e.g. BitTorrent and eMule) to facilitate the delivery of content. In this paper, these Kademlia-based networks deployed by BitTorrent and eMule are called K networks. It is essential for each peer in K networks to possess a unique IP address (or ID), on which both “peer lookup” and “resource searching” rely. However, it is noticed that a significant portion of peers have IP repetition and ID aliasing, through our analysis. In this paper, we propose a couple of metrics to deeply understand the distribution characteristics of IP repetition and ID aliasing. Based on these metrics, we carry out a series of measurement using the P2P crawler named Rainbow. We identify many interesting characteristics of IP repetition and ID aliasing in K networks, which could help promote P2P-network mining.