View all news

Post Deadline Paper accepted at ECOC2017

20 September 2017

The paper entitled ‘Field trial of Machine-Learning assisted and SDN-based Optical Network Planning’ is breaking new ground and is an amazing achievement for the whole Smart Internet Lab team.

Dr Shuangyi Yan 2

This week over 5,000 decision makers from all around the world will come together in Gothenburg, at ECOC, Europe’s largest optical communications exhibition, to gain insights into the industry’s latest developments, new products and trends. A contingent of researchers from the High Performance Networks Research Group, part of the University of Bristol’s Smart Internet Lab, are there all week. We’re delighted to announce that our post deadline paper around machine learning for optical networking, submitted in collaboration with Hong Kong Polytechnic University, has been accepted. Dr. Shuangyi Yan, the lead author of the paper, will present it on Thursday 21 September as part of the conference programme.

The paper discusses an SDN based network planning framework, utilising machine-learning techniques, network configuration and monitoring database, implemented over an optical field-trial testbed comprised of 436.4-km fibre. This demonstrates the adaption of the spectral efficiency utilising probablistic-shaping BVT, based on link performance prediction.  

Dr Shuangyi Yan's commented, "This work shows a real-time network-scale monitoring and configuration database could allow network applications to optimize the whole network dynamically, with data analytics technologies. In the near future, we may open the database running on the NDFIS facility to researchers around the world, to work on our data-driven dynamically optimised SDN network vision together".  The full paper can be read here: Field trial of Machine-Learning assisted and SDN-based Optical Network Planning (PDF, 1,369kB) 

Professor Dimitra Simeonidou, head of High Performance Networks Research Group and Director of the Smart Internet Lab reiterated, “Once again we have proven we are leading the networking groups in the world. I am very proud of my team, of their talents and achievements, and it is inspiring to be able to share our work with world leaders in the field of optical communications”.

To find out more about our work, please contact Maria Korea, our Smart Internet Manager on +44 (0) 117 33 14788 or follow our updates on Twitter @bristol_smart.

Follow twitter updates from the ECOC exhibition by following @ECOC_Exhibition and #ECOC2017.

The article was also featured on the University of Bristol news feed, 22/09/17. 

Further information

About the Smart Internet Lab 

The Smart Internet Lab builds on strategic and applied research on key communications and digital technologies at the University of Bristol to create, and permanently support, a hub for Internet Research with long lasting benefits for society and economy.

About the High Performance Networks Group (HPN) 

The High Performance Networks Group (HPN) led by Professor Dimitra Simeonidou specialises in the application of advanced hardware and software technologies, targeting the future optical networks for Internet of Things (IoTs), data centres, grid/cloud-based applications and distributed technologies etc.  The group is equipped with world-class laboratories, including state-of-the-art optical transmission testbeds, and software-defined network experimental platform.

The group has been at the forefront of world research and development in the fields of:

  • Next generation optical transmission networks;
  • Optical packet and burst switching;
  • Optical data centre solutions and architecture;
  • Grid and cloud networking;
  • Software-defined networking (SDN) and optical network virtualization;
  • Hybrid-network domains orchestration and service management;
  • Smart City ICT Solutions.

The HPN laboratory has dedicated connectivity to both national and international research network infrastructures and forms part of the UK National Dark Fibre Infrastructure Service (NDFIS). This network infrastructure has enabled the development of close collaborations with leading research institutions and industry across the UK, Europe, USA, Brazil and Japan.

Edit this page