Online P2P file sharing networks rely on the cooperation of participants to function effectively. Agents upand download files to one another without the need for any central authority. If agents all contribute to the network and share roughly the same amounts of data as they contribute the network will operate, however if some agents decide to defect and consume far more resources than they contribute the file sharing will stagnate. In online networks with some kind of central authority, such as Ebay, Airbnb, etc. cooperation is achieved through a review system, which is maintained and secured by the central authority. P2P networks are however distributed and cooperation must be achieved without this central mitigator. One way of approaching this problem is by observing cooperative biological communities in nature. One finds that cooperation among biological organisms is achieved through a mechanism called indirect reciprocity. Indirect reciprocity is based on a reputation scheme in which agents share information about each other’s cooperativeness aiding one another in deciding who to interact with and who to shun. In this work we analyse properties a reputation mechanism must satisfy in order to achieve cooperation in P2P networks, incentivising contributions and penalising excessive comsumption of data. In particular, we determine under what conditions reputation mechanisms are resistant to attacks on the P2P network. We focus on one attack above all, namely that of a sybil attack in which a malicious agent creates multiple fake identities who report high levels of cooperativeness about one another. We determine properties accounting mechanisms must satisfy in order to prevent attackers from obtaining arbitrarily high reputation and to consequently be able to consume arbitrarily large amounts of data. This thesis offers a theoretical framework for evaluating the effectiveness of reputation mechanisms on the basis of their ability to induce cooperation and their resistance to sybil attacks.