P2P Networks with IP Based Communication

Survey on Existing Peer-to-Peer Reputation-Based Systems

The peers in the P2P network have to be discouraged from leeching on the network. It has been shown in Tragedy of Commons that a system where peers work only for selfish interests while breaking the rules decays to death. Policing these networks is extremely difficult due to the decentralized and ad hoc nature of these networks. Besides, P2P networks, like the Internet, are physically spread across geographic boundaries and hence are subject to variable laws.

The traditional mechanisms for generating trust and protecting client-server networks cannot be used for pure P2P networks. This is because the trusted central authority used in the traditional client-server networks is absent in P2P networks. Introduction of a central trusted authority like a Certificate Authority (CA) can reduce the difficulty of securing P2P networks. The major disadvantage of the centralized approach is, if the central authority turns malicious, the network will become vulnerable. In the absence of any central authority, repository, or global information, there is no silver bullet for securing P2P networks.

P2P Networks

P2P networks can be categorized into two types:

• •

  Structured P2P Networks;

• •

  Unstructured P2P Networks.

In structured networks, all the nodes know the fundamental structure of the network and hence can prune their search to the relevant nodes. In unstructured networks, there is no relationship between the data or metadata and its location. The unstructured P2P networks do not have a well known architecture.

The reputation schemes proposed in this paper are independent of the structure of the network. As a result, the proposed scheme is equally useful in both the unstructured and the structured networks. The knowledge of the structure of the network can be used for optimizing the algorithm. We do not assume any such knowledge in this paper.