Abstract

Virtualization is instigating a paradigm shift in the networking industry, to keep up with emerging application’s quality of service requirements, massive growth in traffic volume, and to reduce capital and operational expenditures. Network virtualization coupled with function virtualization enables network providers to offer on-demand virtualized networks and services. Network slicing goes a step further by facilitating a new business model, namely Network-as-a-Service, to offer dedicated and customized network slices (i.e. partitions of physical infrastructure) to multiple tenants, while ensuring proper isolation. However, this shift introduces new challenges for network providers and calls for intelligent and automated management. Artificial intelligence and machine learning are considered as enablers for the automated deployment and management of virtualized networks and services. This chapter exposes the state-of-the-art research that leverages artificial intelligence and machine learning to address complex problems in deploying and managing virtualized networks and services. It also delineates open, prominent research challenges and opportunities to realize automated management of virtualized networks and services.