On the Evidence Based Peer-to-peer Resource Management in Distributed Computing Systems

In this paper we propose an accountability management framework for resource sharing in peer-to-peer (P2P) distributed computing systems. On the basis of a generalized e-coin paradigm, we developed the management framework based on the distribution of resource credit/policy, exchange of service and service evidences, and accountability management. A tree topology is assumed for the distribution of resource credit and policies, so that P2P parties can have an absolute reference of the credential of principals, and make the system scalable. Common resource management operations involving splitting, merging and the propagation of system attributes (such as the ownership of the resources) are bound to the resource credit to become resource evidence after the services are rendered. We propose three e-coin binding techniques for those resource delegation scenarios, so that the credit policy (distributed by the credential authority), the resource attribute (sent from the provider) and the resource order (acknowledged from the consumer) can be cryptographically tied with the warrant, the response and the challenge respectively in the e-coin for the generation of the anonymous evidence on the delegation events. Index terms – peer-to-peer computing, accountable resource sharing, e-coin, attribute binding. 1 This work is supported in part by an NSF ITR grant, EIA-0081761, and by a TAMU-CONACYT grant 2 Correspondence author