Intelligent agents in network and service management

This paper proposes a new approach for constructing intelligent agents in the context of network and service management, which is based on, and extends, the ITU-T TMN framework. This knowledge-based approach for intelligent agents in management uses the semantic extensions of network management information for knowledge representation. Different from the current behaviour-oriented semantic frameworks, which do not find wide acceptance in the area of network and service management, a context driven approach is proposed which supports the direct derivation of management decisions. The semantics are defined by context-specific restrictions represented as logical relationships. By structuring the situation descriptions in a hierarchy, we build up a semantic framework for the network and service management which corresponds to the object-oriented development style and benefits from features like reusability, abstraction support, and accumulative developments.