Long-term forecasting of optical Quality of Transmission

The purpose of this project is to investigate long-term forecasting of optical quality of transmission (QoT) based on deep neural networks (DNNs). QoT refers to the properties of the optical channel such as signal-to-noise ratio (SNR), which define the level of service provided by the optical channel. DNNs are a compelling technology, but their usability beyond scenarios outside of image, speech or text pattern recognition is mostly unproven. In the forecasting context DNNs have been used succesfuly for short-term forecasts, which are not useful for management of optical networks. As the scope of use of DNNs is fairly limited at the moment, we plan to investigate new DNN architectures for the purpose of long-term forecasts in optical network management.
In this project, we intend to: (1) investigate the use of DNNs on field data that Ciena has collected from its customers’ networks and evaluate the ability of DNNs to forecast QoT far into the future and (2) develop algorithms that enable estimates of QoT relevant to how service level agreements (SLAs) are specified.

Faculty Supervisor:

Christine Tremblay

Student:

Partner:

Ciena Corporation (St-Laurent, QC)

Discipline:

Engineering

Sector:

Information and cultural industries; Manufacturing; Professional, scientific and technical services

University:

École de technologie supérieure

Program: