Scalable Resource Control in Active Networks

The increased complexity of the service model relative to store-and-forward routers has made resource management one of the paramount concerns in active networking research and engineering. In this paper, we address two major challenges in scaling resource management to many-node active networks. The rst is the use of market mechanisms and trading amongst nodes and programs with varying degrees of competition and cooperation to provide a scalable approach to managing active network resources. The second is the use of a trust-management architecture to ensure that the participants in the resource management marketplace have a policy-driven \rule of law" in which marketplace decisions can be made and relied upon. We have used lottery scheduling and the Keynote trust-management system for our implementation, for which we provide some initial performance indications.