Collusion in peer-to-peer systems

Peer-to-peer systems have reached a widespread use, ranging from academic and industrial applications to home entertainment. The key advantage of this paradigm lies in its scalability and flexibility, consequences of the participants sharing their resources for the common welfare. Security in such systems is a desirable goal. For example, when mission-critical operations or bank transactions are involved, their effectiveness strongly depends on the perception that users have about the system dependability and trustworthiness. A major threat to the security of these systems is the phenomenon of collusion. Peers can be selfish colluders, when they try to fool the system to gain unfair advantages over other peers, or malicious, when their purpose is to subvert the system or disturb other users. The problem, however, has received so far only a marginal attention by the research community. While several solutions exist to counter attacks in peer-to-peer systems, very few of them are meant to directly counter colluders and their attacks. Reputation, micro-payments, and concepts of game theory are currently used as the main means to obtain fairness in the usage of the resources. Our goal is to provide an overview of the topic by examining the key issues involved. We measure the relevance of the problem in the current literature and the effectiveness of existing philosophies against it, to suggest fruitful directions in the further development of the field.