Danish Rafique, ADVA Optical Networking, Germany;
Antonio Napoli, Coriant R&D GmbH, Germany
; Yawei Yin, Alibaba Group, China
Artificial Intelligence (AI) and Machine Learning (ML) are ubiquitously emerging at different levels within optical networks. This workshop aims at discussing the potential needs and advantages as well as limitations and risks of AI-driven solutions for a number of use cases. This will include predictive maintenance, impairment estimation, control plane automation and related security issues.
For example, how much benefit do we expect in terms of quality of transmission (QoT) estimation? Can AI effectively boost the way today’s network are managed towards fully automated operations, or shall we just expect relatively minor improvements in localized management tools?
Both data center and optical transport networks viewpoints will be discussed, aiming at deployment status and expected potential to optimize network resource usage, introduce proactive solutions and ultimately reduce OPEX costs. What kind of AI-based solutions will be deployed in the short and long term? Which use cases should be prioritized, if at all? Are the system vendors and network operators ready to delegate AI, and thus reduce their control and security, to perform fully automated network operations?