Abstract:

Modern system management needs the capabilities to deal with a combination of areas, ranging from administration of elements to provision of services and management of the enterprise itself. As such, traditional element-based network management views are being rapidly replaced by integrated management approaches that take the nature of the enterprises and the services they provide into account. More conventional computer applications and established protocols, like Simple Network Management Protocol, provide automation to some extent. But generally speaking, human interactions are crucial for their proper operations. The need for such interactions mainly stems from the incoherency, incompleteness, and conflicting data with varying degrees of relevance available for achieving the management functions. In such complex environments, artificial intelligence-based solutions can be utilized to improve integrated management efficiency. This paper elaborates on these topics. In particular, this work describes several ways that fuzzy representations and knowledge granulations can be used to identify or to improve the solutions to problems encountered in an integrated network management environment. Some specific application areas that demonstrate the effectiveness of these ideas in improving management of the networks are also discussed.