An Intelligent-agent-based Architecture for the Management of Heterogeneous Networks

The management of heterogeneous networks requires the capabilities to combine differing types of data and to account for events occurring on different time scales. No single network management methodology is likely to satisfy the diverse needs of today’s networks so as to maintain a desirable level of quality-of-service (QoS)–based performance. Rather, an integrated approach is needed in which a combination of methodologies can respond to events at different time scales, which can occur in large heterogeneous networks. In this paper, we present our ongoing work to address this need: we are developing a framework that can provide distributed, adaptive, and automated network management for today’s complex networks. Our model is based on the use of intelligent agents that use semantic and statistical models to correlate and analyze data; control decisions are made both locally and globally from a QoS perspective. A simulation model of a wireless network, also in development, will utilize and test our approach to network management.