Optical Networks for the Next Decade
The volume of user traffic on carrier networks is growing at between 20% and 40% per year and if this trend continues as expected then optical transport networks will become overstretched without a significant increase in capacity beyond that currently planned by operators.
The solution to this problem is likely to include a mix of new technologies, expanded bandwidths, network autonomy and many more fiber links, up to 10x in the next ten years. Understanding and solving these challenges is the key focus of the IEEE-SA Optical Networks 2020 (ON2020) group that includes individuals from leading carriers, equipment manufacturers and component vendors.
During a workshop on the show floor at OFC 2019 in San Diego this month, the ON2020 group presented some initial findings and gave an insight into the challenges the industry faces and the ongoing work within ON2020 to understand these challenges and identify solutions.
The workshop started with a short presentation by Rudi Schubert, Director of New Initiatives at the IEEE, who introduced ON2020 and the IEEE-SA Industry Connections Program. Peter Winzer, chair of ON2020, then gave an overview of the goals for the initaitive and introduced the six key areas on which discussion groups within ON2020 are focusing. These areas are: traffic evolution in optical networks; optics integration onto switch engines; optics integration onto coherent engines; transmission systems; transport network protocols; and network autonomy and control.
Albert Rafel from BT gave the first group presentation, covering traffic evolution, on behalf of Andrew Lord, the discussion group leader. Rafel explained there is some evidence that the exponential traffic growth is slowing down but that the introduction of 5G may introduce new use cases. The group's very tentative conclusions are that capacity growth will continue as it has, and this may mean that 10x more fiber links will be needed in ten years if the industry sticks with the conventional C-band.
Chris Cole from Finisar gave an overview of the problems that could be solved by the integration of optics onto switch engines, identified some of the candidate technologies and architectures that are being discussed, and detailed the technical challenges that need to be addressed. Cole explained he is expecting this to be the most exciting topic and is anticipating significant participation in the discussions within the group.
The next presentation was given by Winston Way from NeoPhotonics, who is leading the discussions about optics integration onto coherent engines. This group had already received contributions from 17 experts representing 13 different companies. Way shared many views on the maximum rates for single wavelength coherent solutions, the development of coherent DSPs, the limitations on the use of superchannels, and the material and packaging options for optics integration onto coherent engines. He also shared views on the use of coherent technology by 2023 for intra- and inter-data center interconnect.
Presentations given by Sethumadhavan Chandrasekhar from Nokia Bell Labs and Maarten Vissers from Huawei covered transmission systems and transport network protocols. The transmission systems group is looking at the evolution of transmission systems over the next five to ten years, including: bandwidth expansion beyond the C-band to include L- and S-bands; the evolution of ROADM/OXCs; superchannels; maximizing capacity of undersea transport systems; and the implications for fiber deployments of expanded space division multiplexing.
The transport network protocols group is addressing the changes to network protocols as the optical network changes from today's network, based on hardware layers, to one where a software-defined optical service network (OSN) is provided on top of optical and electrical layers. This software-defined optical network will support SDN slices, including SDN Slice as a Service (SDNSaaS) and Frequency and Time Synchronization as a Service (FTSaaS).
The final presentation, given by Martin Burke from AT&T, covered network autonomy and control. The network autonomy and control group expect this new optical network to be managed using data-driven decisions and policy with machine learning (ML) support, enabling a high level of autonomy and minimizing human intervention. The discussions in this group are around the challenges and opportunities this approach presents.
Together these six groups within ON2020 are addressing the key challenges facing the industry in ensuring the optical transport network is ready to handle the expected traffic demands from users during the next five to ten years. ON2020 is planning further activities during 2019, including a full readout of results and roadmap discussions at IEEE Globecom in Hawaii during December.
If you wish to join the discussion in ON2020 or find out more, please sign up through IEEE-SA - Optical Networks 2020 Program.
This article was sponsored by ON2020
— Simon Stanley, Analyst at Large, Heavy Reading