A scalable hierarchical architecture for distributed network management

Abstract

We present a hierarchical, scalable and dynamic management scheme for managing large-scale distributed networks. The proposed hierarchical architecture achieves scalability by modifying the number of components according to the size of the network. The hierarchical approach can efficiently distribute the loads and improve the reliability of the system. In order to increase flexibility, the system employs a policy-based dynamic subscription approach that enables users to manipulate management tasks at run-time. The event-triggered management mechanism enables the managers to discover and solve the problem of network in time. Moreover, the system can trigger or disable management rules automatically according to some network events. This significantly improves autonomy of the system.