Workshops

Workshops will be held on Sunday, 20 September 2020, and are open to all conference registrants.
Each workshop will have a total of two sessions of two hours each, including a panel for interactive discussions. The format of each session is determined by the organizers.

Current list of scheduled workshops:

  • WS 1: Ultra-wideband systems: next frontier of research
  • WS 2: The technologies for next-generation ultra-high-density cables - New cable structure, new fibers and new relating technologies
  • WS 3: If, how, and where is artificial intelligence (AI) beneficial in optical networks and interconnects?
  • WS 4: Co-Packaging of Photonics & Electronics
  • WS 5: Photonics-Electronics Integration: Opportunities and Challenges for Signal Processing
  • WS 6: Pathway to Bring Photonics in High-Performance Computing: from Materials to Applications
  • WS 7: Fabless start-ups developing PIC-based products: opportunities and challenges
  • WS 8: Free-space and deep-space optical communications
  • WS 9: Constellation shaping – All set for maximizing capacity?
  • WS 10: Advances towards Millimeter- and Terahertz-wave antenna arrays for broadband wireless
  • WS 11: Coherent optics for Intra-Data Center/Campus Optical Interconnects
  • WS 12: Is the time ripe for ultra-high bandwidth photonics in data centers?
  • WS 13: Post 5G edge cloud computing: why, what and when?
  • WS 14: The next generation of fixed networks by ETSI F5G (currently scheduled on Tuesday 22 September)

 

More detailed & updated information on Sunday workshops:

WS 1: Ultra-wideband systems: next frontier of research

Organisers: Uiara Celine di Moura (DTU Fotonik) and Maria Ionescu (Nokia-Bell labs)

Abstract: Ultra-wideband (UWB) systems are emerging as a viable solution to cope with the exploding capacity demands in optical networks. These systems leverage the already deployed single mode fibers by extending the transmission over the low-loss optical spectrum beyond the conventional band. This approach could potentially be a more readily available and cost-effective solution to increasing capacity, when compared to space division multiplexing (SDM) approaches. UWB systems could additionally enable a pay-as-you-grow strategy, allowing operators to add more bands as needed. However, the deployment of UWB systems also comes with a set of challenges, such as the availability of optical components, physical integration, system management, optical power optimization, bandwidth allocation, and electrical power consumption. The aim of this workshop is to present ongoing research addressing some of these design challenges and foresee future developments in UWB systems.

 

WS 2: The technologies for next-generation ultra-high-density cables - New cable structure, new fibers and new relating technologies

Organisers: Kazuhide Nakajima (NTT), Pierre Sillard (Prysmian Group) and (TBC) Durgesh Vaidya (OFS)

Abstract: Space division multiplexing (SDM) is recognized as a key technology to overcome the future capacity crunch. The fibre count in an optical cable has been rapidly increasing in these years. Optical submarine network operators have already considered 12 optical fibre pairs in their next systems, and further more fibre pairs are expected to be implemented in future systems. In a terrestrial network, thinner coating and/or rollable optical fibre ribbon enable us to realize a high-density optical fibre cable, and an optical cable which contains 1,000 to 6,912 optical fibres is expected be beneficial for inter DC application. To accommodate more than 10,000 optical fibres in single small-diameter optical cable or to realize higher density cable, this workshop discusses what kind of cable structure, what kind of fiber, and what kind of relating technologies would be required.

WS 3: If, how, and where is artificial intelligence (AI) beneficial in optical networks and interconnects?

Organisers: Lena Wosinska (Chalmers Univ. of Technology), Francesca Parmigiani (Microsoft), and Chongjin Xie (Alibaba)

Abstract: In the recent years AI technologies have been proposed and claimed to be instrumental across a variety of industrial sectors, including optical networks. Their applications in both optical networks and data center interconnects have been widely investigated in all the various layers (physical, network and service), for example, to increase fiber link capacity, reduce operational cost, optimize resource usage, and support new services. In this workshop we will discuss the real needs and performance advantages as well as the associated challenges of AI-driven solutions for a number of use cases in both optical networks and data center interconnects.

The workshop will address the following topics:

  • • New trends in optical networks control and automations: in what extent and at what pace will AI be introduced in network operation? What are the benefits and what are the challenges?
  • • Where should AI be used, data plane, control plane or both?
  • • What are the advantages of using AI in optical networks and interconnects over conventional techniques based on modelling and theory?
  • • How to use AI in performance monitoring and how it can improve the optical network agility?
  • • Can AI improve the optical networks efficiency in terms of resource utilization, energy consumption and reliability performance?

The workshop will be divided in two parts. The first part will consist of short presentations (15 min incl. QoA) given by the experts on the topics targeted by the workshop. The second part will be an interactive panel discussion involving both the experts and the audience.

WS 4: Co-Packaging of Photonics & Electronics

Organisers: Peter O’Bien (Tyndall Institute), and Thomas Liljeberg (Intel)

Abstract: not available yet

WS 5: Photonics-Electronics Integration: Opportunities and Challenges for Signal Processing

Organisers: Xi Chen (Nokia Bell-labs), Stephan Pachnicke (Kiel Univ.) & Christoph Scheytt (Paderborn Univ.)

Abstract: Today optical and electronic circuits are still clearly separated domains. In the recent past, however, photonic-electronic integration is developing rapidly based on either Silicon photonics or Indium-Phosphide technology platforms. This enables close proximity between optics and electronics and consequently higher bandwidths, reduced energy consumption and size. The close integration also opens up new approaches to signal processing both in the analog and digital domains.

This workshop will investigate how photonic-electronic integration can pave the way to completely new system concepts. It will discuss examples of future signal processing circuits and elaborate challenges and limitations.

WS 6: Pathway to Bring Photonics in High-Performance Computing: from Materials to Applications

Organisers: Francesca Parmigiani (Microsoft), Bhavin Shastri (Queens University), Paul. R. Prucnal (Princeton University), Thomas Krauss (York University).

Abstract: In the recent years, microelectronic computers have encountered challenges in meeting all of today’s demands for new applications in high-performance computing. Neural network models have come to dominate modern machine learning algorithms, and combinatorial optimization problems solvers are extremely important in multiple sectors, such as drug discovery and finance.

Continuing meeting these demands, as the size of the network or problem scales, will require the development of unconventional computers employing alternative processing models and new device physics. Dedicated pieces of electronic hardware have been developed to speed up specific operations and implement them more efficiently. 

Photonic integration industry promises to bring manufacturing ecosystems normally reserved for microelectronics to photonics. Photonic devices have already found simple analog signal processing niches where electronics cannot provide sufficient bandwidth and reconfigurability. In order to solve more complex information processing problems, they will have to adopt a processing model that generalizes and scales. Photonic neural networks aim to map physical models of optoelectronic systems to abstract models of neural networks. Similar discussion applies to coherent Ising machines. These field represent new opportunities for extremely-fast and energy-efficient machine information processing, with application to mathematical programming, intelligent radio frequency signal processing, and real-time control.

Ranging from materials, devices, systems, architectures, algorithms, and applications, this workshop will cover topics on the current status of the field of high-performance optical computing, including photonic neural networks, coherent Ising machines and emerging approaches such as quantum optical neural networks and superconducting optoelectronic networks. Future challenges and advances in science and technology to meet challenges will also be discussed. Finally, the workshop will aim to address the challenges associated with commercializing this technology driven by applications domains.

The workshop will be divided in two parts. The first part will consist of short presentations (10 min plus few minutes QoA) given by the experts on the topics targeted by the workshop. The second part will be an interactive panel discussion involving both the experts and the audience. 

WS 7: Fabless start-ups developing PIC-based products: opportunities and challenges

Organisers: Roel Baets (Ghent University – imec)

Abstract: Photonic Integrated Circuits (PICs) have stamped their presence in the optical transceiver market through the launch of various transceiver products by key players of that market. In the last few years, many fabless startups are leading an equally impressive entry of PICs into new markets to offer solutions for sensing, medical diagnostics, LiDAR, 5G, artificial intelligence, and many other applications. PIC technologies and associated services are gaining maturity at an unprecedented pace. This rapidly consolidating ecosystem can fulfill the demands of fabless companies operating in new markets. 
This workshop aims to:
•present how various startups develop innovative products in various new markets based on PICs
•address outstanding challenges faced by fabless companies with respect to the capabilities of the PIC ecosystem

WS 8: Free-space and deep-space optical communications

Organisers: Jade Wang (MIT) + others to be identified

Abstract: Not available yet

WS 9: Constellation shaping – All set for maximizing capacity?

Organisers: Alex Alvarado (TUE) and Tobias Fehenberger (Adva)

Abstract: Geometrical constellation shaping (GCS) and probabilistic constellation shaping (PCS) promise to squeeze the last tenths of a dB out of high-capacity optical fiber links. However, the optical fiber community seems to have different opinions about the future of signal shaping. One camp believes that it is end of the road for GCS and PCS in the sense that only the engineering problem of implementing and productizing well-understood methods is left. The second camp argues that more work needs to be done to fully understand the trade-offs associated with the available techniques, as well as to find ways to unlock the maximum shaping gain of the nonlinear fiber channel.

In this workshop, speakers from academy and industry from both camps will share their views on the future of constellation shaping and the prospects of maximizing capacity by addressing the following questions: Will signal shaping just be an off-the-shelf—possibly even standardized—part of our DSP toolbox? What else can we improve on the DSP side (including FEC, implementation penalties, adaptability) and is it worth doing so in the system context (considering latency, complexity, cost, power consumption)? Is the rate adaptivity offered by PCS a killing argument to completely forget about GCS? How much rate adaptivity do we actually need in optical networks? And why not achieving this via rate-adaptive FEC or time-hybrid modulation formats? On top of constellation shaping, what tools could make it into our DSP toolbox to achieve even higher throughput?

WS 10: Advances towards Millimeter- and Terahertz-wave antenna arrays for broadband wireless

Organisers: Guillermo Carpintero (Universidad Carlos III de Madrid) and Tadao Nagatsuma (Osaka University)

Abstract: Microwaves, Millimeter-wave and Terahertz frequency bands are being pursued for beyond 5G, enabling ultra-broadband wireless communications, to unlock 100 Gb/s data rates. Because these bands all together occupy but a tiny fraction of the spectrum of optical spectrum, photonics technologies can be used to efficiently propagate and manipulate these signals. Actually, photonic technologies have already demonstrated broadband photonic signal generation, up-conversion and down-conversion, as well as true-time-delay photonic beam-forming of antenna phased arrays. This workshop aims to present a broader view of the current state-of-the-art in crucial developments to enable photonic driven beam-forming and beam-steering  antenna phased arrays. This workshop covers integrated microwave photonics approaches, in which photonic integration technology is employed to overcome the energy-efficiency, flexibility and scalability, as well as performance limitations of the current systems using optical fiber and discrete-components. The program includes as well additional key technologies, such as novel antenna approaches and dielectric waveguide structures.

WS 11: Coherent optics for Intra-Data Center/Campus Optical Interconnects

Organisers: Ilya Lyubomirsky (Inphi) and Xiang Zhou (Google)

Abstract: Recent technology trends are showing significant challenges in scaling the IM-DD PAM architecture for 2-10km intra-campus data center interconnection applications.

Meanwhile, the success of 400G ZR has brought coherent pluggable module technology into the 10-100 km space, providing the lowest cost/power solution for >10km DCI applications.  We expect coherent optics to similarly disrupt intra-data center and intra-campus applications for reaches below 10km. However, intra-data center/campus networks are very different from telecom DWDM networks, and include a rich variety of short reach optical technologies, including VCSEL MMF for < 100m, parallel SMF systems for ~ 500m, and CWDM SMF systems for 2km “FR4” and 10km “LR4” applications.

This workshop will focus on the intersection of coherent technology with next generation 800G and 1.6T optical interconnects for 2km and 10km applications. The following questions will be addressed:

  1. How does coherent optics compare with IM-DD PAM and in which applications would coherent technology have the advantage?
  2. What innovations in optics are necessary to achieve the lowest cost?
  3. What is the optimum QAM modulation format? Is it different for 800G vs. 1.6T?
  4. What innovations in DSP algorithms are necessary to drive down power consumption?
  5. What is the right FEC to provide sufficient NCG for coherent systems while keeping latency < few hundred ns?
  6. How to address the backward compatibility challenge facing coherent optics ?

WS 12: Is the time ripe for ultra-high bandwidth photonics in data centers?

Organisers: Ioannis Tomkos (Univ. of Patras) and Yvan Pointurier (Nokia Bell-labs)

Abstract: As many stakeholders predict the end Moore's law for electronic switching in data centers, and server interface and switching port capacity keep increasing, optics is increasingly leveraged and stressed. Of course optical components offer larger bandwidths than electronic components, but they are also limited by the underlying electronics - for instance drivers and analog/digital converters. Will optics suffer the same fate as electronics and inherently limit data center scale, or will they enable new paradigms in data centers?

Our mix of academic and industrial speakers will tackle this question, focusing on the key infrastructure aspects of data centers. The speakers will review the current limitations faced by current data centers and architectures, and will discuss whether co-packaged optics is the way forward to circumvent Moore's law or just a fad. On the transmission side, they will introduce new components that could replace current pluggables to reach unseen short reach transmission capacity. Will optical (analog) signal processing, a hype in the 1990s before the disillusion of the 2000s, make a come-back in the 2020s to alleviate the pressure on electronic processing? Our speakers also have a good idea on that. And finally, the speakers will discuss the future of optical switching in data centers.

WS 13: Post 5G edge cloud computing: why, what and when?

Organisers: Sébastien Bigo (Nokia Bell-labs) and Nicola Calabretta (TUE)

Abstract: Originally motivated by prospects of cost savings from virtualizing network functions of 5G networks, the distributed information technology (IT) infrastructure of edge-clouds comes with unique advantages, which will undoubtedly attract digital service providers into new business opportunities. One can easily predict that the largest potential for revenue creation will leverage its unmatched responsiveness. Overall, even if this new IT infrastructure is yet to be fully agreed upon, it will undoubtedly leverage highly-distributed IT resources, either ideally located in the 5G sites, or part of the thousands of new small data centers (DCs) being erected by cloud operators, as they move away from today’s dominant model of central cloud - where IT jobs are performed inside single location DCs. Regrettably, distributing IT over could steer the profitability of cloud businesses down because serves are more poorly utilized in average, unless machines become capable of efficiently cooperating across multiple locations. The workshop is meant to be a forum for major stakeholders to exchange on challenges and opportunities for this promising new paradigm of computing.

WS 14: The next generation of fixed networks by ETSI F5G (scheduled on Tuesday 22 September)

Organisers: Wangyin James(Huawei)

Abstract: not available yet