Semantic-Based Decision Support for Network Management and Orchestration

Software-Defined Networking exploits Network Function Virtualization to face the challenges of modern enterprise networks, integrating cloud and Internet of Things deployments. Conventional methodologies are often inadequate, whereas novel approaches based on Artificial Intelligence can achieve the required levels of flexibility and scalability. This paper presents a network orchestration and management framework based on SemanticWeb languages and technologies. The proposal adopts a two-level ontology design: (i) a novel translator automatically generates a low-layer ontology from the YANG network information model; (ii) a high-level domain ontology is introduced to model the contextual scenario knowledge. In early tests, a prototypical implementation of the proposed framework has exploited SWRL rules to check network requirements against metrics and constraints, in order to detect violations and provide context-aware suggestions for corrective actions.