O2. Machine learning for optical network and transmission - why and where?

Machine Learning (ML) technologies are emerging and ubiquitously spread all over the industries including optical networks. Their applications in both optical transmission and optical networking have been widely investigated in recent years to increase fiber link capacity, reduce operational cost, optimize resource usage, and support new services. This workshop aims at discussing and clarifying the potential needs and advantages as well as limitations and risks (e.g. reliability issue caused by insufficient and/or inappropriate training) of ML-driven solutions for a number of use cases in optical transmissions and networks. This will be expected to facilitate innovation in aspects such as channel impairment compensation, end-to-end link optimization, physical layer monitoring and the autonomous control and management of future optical infrastructure.

Takahito Tanimura
Fujitsu Laboratories Ltd.
Qunbi Zhuge
Shanghai Jiao Tong University
O2+O3. Recent efforts and progress towards real deployment and standardization of SDM technology

Space division multiplexing (SDM) in the form of multi-core fibers (MCF) is a potential technology for future optical transmission systems and has been subject of intensive research over the last decade. MCF has successively demonstrated transmission speeds in excess of Petabit/s, however, with regard to actual deployment, there still remain a number of unanswered questions and unsolved problems, both from a technical and commercial point of view. This workshop is an excellent opportunity for industry’s experts to address these open questions, which could eventually pave the way towards standardization and commercialization of MCF-based solutions.
The focus of this workshop is to investigate the practical elements associated with an end-to-end MCF ecosystem. More specifically, the workshop will explore the problem statements that MCF technology is better suited to address relative to single-core fiber technology, applications spaces where MCF is more likely to occur first, MCF designs that are most pragmatic, MCF fiber/connectivity performance requirements, achievable cost-per-bit of MCF-based solutions relative to its single-core fiber counterparts, and factors that will either hinder or accelerate MCF technology adoption. We would also encourage the authors to explore on whether alternative techniques (such as, few-mode fibers) could make MCF unattractive.

Werner Klaus
Jing Li
Yangtze Optical Fibre and Cable Joint Stock Limited Company
Peter Pondillo
Corning Incorporated
Takehiro Tsuritani
KDDI Research, Inc.
O4+O5+P1. How to establish a sustainable ecosystem for photonic integrated circuits? What are major hurdles to overcome?

Photonic integrated circuits (PICs), which have been designed and fabricated on III-V and silicon-photonics platforms including photonic-electronic integration with peripheral electronic circuits and optical input/output interfaces, have enabled the emergence of versatile photonic components of small form factor with high energy efficiency. The photonic components have been distributed extensively in photonic networks for high-capacity communication and high-performance computing, and have brought the evolution to our community. Foundries and enterprises to conduct the manufacturing of PICs have been established, thereby small-form-factor energy-efficient photonic components can be manufactured with precision and reliability in fabrication processes. For further enhancement in the manufacturing of PICs in terms of lead time, cost, and carbon footprint without deteriorating the precision and reliability, there are growing demands toward an ecosystem to integrate supply and service chains of design, fabrication, and characterization of PICs based on the photonic-electronic integration platforms. In this Joint Workshop, current status and challenges on ecosystem development for PICs are reviewed through presentations given by prominent leaders in the forefront of practical applications of PICs from the points of view of ecosystem constructors and users. Intensive discussion is also organized in panel session following the presentations to gain insights into overcoming obstacles against the ecosystem development and envisage our path to a sustainable ecosystem for PICs.

Giampiero Contestabile
Scuola Superiore Sant'Anna
Yoshiko Inoue
Sumitomo corporation Kyusyu
Kensuke Ogawa
Tokyo Institute of Technology
P2. Wired and wireless network convergence in 5G/IoT era

Seamless convergence between wired and wireless systems is indispensable for 5G/IoT networks by the development of not only advanced wireless systems but also the innovation of the wired networks. In particular, to attain the features on high throughput, low latency, and huge capacity on connected devices, stringent and flexible management of quality of services is demanded in the network side. In the scenario, software-defined network and network function virtualization will play important roles in the resource management as well as slicing multi-services on the network. PHY and L2 layers should be performed for energy efficient and low footprint subsystems like a highly efficient switch. This workshop discusses the function of the optical and network technologies in future wireless access networks.

Atsushi Kanno
Hideki Tode
Osaka Prefecture University