ARE: Action Recommendation Engine for Network Operations Centers (Phase IIb)

点击翻译按钮获取中文摘要

The goal of this project is to provide evidence of real-world deployability and transfer the technology of an invention by the University of Ottawa's research team to the multinational telecommunications company Ciena Corp. (hereafter Ciena) which has significant business and R&D operations in Canada. The research prototype of the invention, called Action Recommendation Engine (ARE), was developed from October 2018 to June 2021 under a $287,635.43 MITACS Accelerate Cluster research project in partnership with Ciena, and is now ready and successfully verified in a lab setting. ARE is an Artificial Intelligence (AI) based system that (1) automatically detects problems in IP networks, (2) gives recommendations on how to fix those problems, and (3) gives recommendations on taking actions to minimize network provider's cost while maximizing users' quality of experience (QoE). ARE can be deployed either as a recommender, requiring a human operator to apply/override the recommended actions, or as a fully autonomous action taker in a human-out-of-the-loop fashion. ARE's applications are mainly in Network Operations Centers (NOC), where ARE can lead to a significant reduction in operation costs by automating some or all of NOC's manual operationsThe objectives of this project are to finetune ARE for a real environment and to integrate it inside a realistic tool such as the company's Blue PlanetTM Intelligent Automation product line in order to validate ARE's design and implementation in a practical customer setting. At the end of this project, we will have finetuned the ARE technology and established evidence of its technical readiness and real-world deployability, while Ciena will have acquired the technical capability to take ARE to the market. Ciena has exclusively licensed ARE and has also patented it.