A P2P-Based Framework for Distributed Network Management

In this paper we present a novel framework supporting distributed network management using a self-organizing peer-to-peer overlay network. The overlay consists of several Distributed Network Agents which can perform distributed tests and distributed monitoring for fault and performance management. In that way, the concept is able to overcome disadvantages that come along with a central management unit like scalability and reliability. The self-organization of the overlay is achieved by the application of a Distributed Hash Table (DHT) based on the Kademlia algorithm. The DHT mechanism implements a distributed index for the rapid localization of other agents or resources. The framework provides a reliable and scalable basis for distributed tests, like e.g. the identification of performance degradation in IP networks using throughput statistics. The framework is not intended to replace the central network manager, but rather to support it in testing and surveying the status of the corresponding network.