Although P2P has proven itself as a viable architectural paradigm for distributed applications, the success of future P2P applications ultimately depends on convincing users to volunteer these resources. In this project, two inter-related P2P incentive paradigms are under investigation. In the first paradigm, called Networked Asynchronous Bilateral Trading (NABT), each user has a set of online friends. Users can trade asynchronously with direct friends using local currency and a debt limits. NABT also allows trades to pass through intermediaries. For file sharing, NABT is almost as efficient as a perfect economy, where all users can trade directly with each other. Research is also aimed at extending NABT, developing designs and theories for heterogeneous P2P resource markets. The second paradigm, called Closed P2P Communities, uses a lightweight centralized banking infrastructure. Here the focus is on designing and analyzing a new class of powerful, but lightweight currency-based incentive schemes. Game theory is being applied to design incentive mechanisms that optimize deal throughput for diverse P2P markets and pricing theory is used to study the pricing dynamics in closed P2P communities. Novel approaches for detecting and evicting colluders are also under development. Prototypes for both incentive paradigms will be constructed. The project will provide a framework for a new P2P computing paradigm. It will involve undergraduate and graduate students, especially minority students. Interactions with industry will be facilitated through the CATT center at NYU-Poly. Educational material will be developed and disseminated for undergraduate and graduate level networking courses taught by the PIs at NYU-Poly.
