Soft-security to self-preserve autonomic systems: on the application of reputation management schemes
In this talk, we first analyze the principles for an autonomic system to discuss new security requirements and issues that target malicious and selfish nodes. We then dissect a reputation management scheme to study the cost vs. the benefit of its application in P2P systems. Borrowing techniques from game theory, we present a framework based on the Iterated Prisoner’s Dilemma to justify the adoption of reputation management schemes in self-organizing networks. We model the interactions of rational and selfish nodes in distributed systems and we study how a node takes into account the change of its reputation when deciding its behavior in a transaction. Finally, we discuss the Nash equilibrium in the system.
Once we have analyzed the importance of reputation management schemes, we estimate their performance for specific applications. We first simulate the capability of these schemes in reducing corrupted file transfers in end-user collaborative content distribution systems. Then, we define a token-based mechanism to improve the application of reputation management schemes in mobile scenarios, i.e., to reduce the problem of reputation bootstrap in new communities and accelerate the identification of malicious nodes.