Modern optical transport network is a highly complex object and introduces a challenge to managing and exchanging of management information in an accurate and recognize manner. Different optical technologies and service delivering rules increase the complexity of the network management activities. This paper analyzes machine-learning techniques of the optical network management with taking into account the knowledge representation based on the object-oriented ontology. Object-oriented ontology can be defined in general as a formal specification of a machine-readable vocabulary which is result of experts analyze. Object-oriented ontology is formulated in the terms of classes or entities and associations between them, to describe knowledge about the contents of the optical transport network management and another related object as quality of transmission, wavelength routing, flexible grid. Object-oriented ontology can be mapped for a machine learning models in the context of general management networks areas as fault management, performance management, configuration management, security management, business process management. Some machine-learning techniques for the optical network test and management are examined. Main entities of optical transport network management and their relations are discussed.
|