This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).
Communication protocols on the Internet are typically designed first to work, then to interoperate and evolve, and eventually to be resilient to misbehavior and selfishness. Unfortunately, these extensions to protocols often sacrifice backward compatibility, because otherwise a misbehaving user can claim to be legacy. At the other extreme, the ``incentive compatible'' protocols are often designed assuming selfish users are also myopic, focusing narrowly on the short term and unwilling to donate resources for no immediate gain. This project investigates the iOwe, a new primitive for exchanging value in competitive protocols, and applies it to two new overlay network protocols, PeerWise and HoodNets. The iOwe is a verifiable promise of future service that can be exchanged by nodes that trust the issuer. PeerWise and HoodNets are protocols by which users can achieve better network performance, either in terms of latency (PeerWise) or wireless bandwidth (HoodNets). Both rely on a non-simultaneous exchange of service: PeerWise because simultaneous use is unlikely, HoodNets because channel bonding achieves better apparent bandwidth only when both links aren't already saturated. The project results are expected to provide new facilities to help build protocols in which repeated interactions permit users to operate at a short term loss for long term gain. Similarly, the protocols being prototyped can aid in the robustness and performance of network service.
