Consistency of states of management data in P2P-based autonomic network management

Abstract

Autonomic network management is a vision that brings autonomic computing principles to network management. Besides, it is necessary some level of decentralization to enable broad autonomic capabilities. An interesting alternative of infrastructure for this union is the utilization of peer-to-peer (P2P) overlays. However, the consistency of states of management data among peers is an important challenge. Traditional mechanisms to maintain consistency of these states are supported by some centralization which wastes some desirable properties of P2P facilities. In contrast to these mechanisms, we propose a distributed, scalable and robust mechanism to maintain the consistency of states of management data. In this paper we introduce multi-agent truth maintenance features in P2P-based autonomic network management as a mechanism to bring consistency maintenance of these states. We developed a model of a P2P-based autonomic network management system on Peersim to perform simulation experiments. Besides, the utilization of P2P-based autonomic network management systems in access networks is presented as a case study. © 2009 Springer-Verlag.