Scalable and self-x optical switching for next generation data centres
1 October 2019
Using optical switching technologies and artificial intelligence methods to design and control agile Data Centres
Funder EPSRC & HUBER+SUHNER Polatis CASE
Amount £ 125 000
Research topics Artificial intelligence and machine learning for device and network optimization and control | ultra fast optical switching | network resource allocation | compute and network resource managment | combinatorial problems
Data Centres of different sizes form the backbone of the Cloud, Edge and Mobile Computing. They shape the digital infrastructure for next generation applications and services across Internet of Things, 5G, artificial intelligence, virtual and augmented reality as well as Tactile Internet among others.
The Ph.D. project investigates optical switching technologies, systems and networks to support next-generation Data Centres. Of particular focus and interest to be explored are elements on scalable optical switching fabric from single switch to reconfigurable topologies to support millions of endpoints (servers, compute/memory/accelerators).
Machine-learning methods and AI algorithms will be explored to deliver dynamic self-x (calibration, planning, learning, scheduling, and optimizing) systems and networks. The activities will be supervised by the PI, Dr. George Zervas (University College London), as well as industrial supervisor Michael Enrico (Huber-Suhner Polatis).