Abstract

Increasing traffic due to increased bandwidth or the number of users calls for scalable systems, which can be built with peer-to-peer (P2P) mechanisms. Scalability is a key issue for systems that rely on many participants, such as large-scale collaboration system. This paper introduces PeerCollaboration, a fully decentralized P2P collaboration system for documents, which is robust against malicious behavior, provides an efficient content search, and offer mechanisms for distributed control. Typical tasks in PeerCollaboration are searching, retrieving, creating, changing, and maintaining documents in a collaborative manner. Three problem areas are investigated within this system. The first problem focuses on similarity search on top of existing P2P networks and highlights a novel algorithm, which outperforms compared approaches. The second problem deals with incentive schemes, which work with indirect reciprocity. Thus, the novel and robust incentive scheme finds more reciprocities than with compared approaches. The third problem focuses on user-based voting mechanisms.