This video isn’t available to you right now
Login to check your access and watch the full session.
To redeem an access code, first log in.
About the session
Co-organisers Session 1: Francesco Da Ros, DTU, Georg Böcherer, Huawei Technologies, Fotini Karinou, Microsoft Research Ltd.
Co-organisers Session 2: Gabriel Charlet, Huawei France, Marco Bertolini, Nokia Italy, Bernhard Spinnler, Infinera Germany
Abstract session 1
Exponentially increasing data center traffic requires innovative low cost and low power optical transceivers. Recent academic and industrial research is disrupting the digital paradigm, replacing (parts of) the DSP by optical filtering, electrical RF filtering, and by electronic, photonic, and optoelectronic analog neuromorphic processing. In this workshop, we analyze in depth the analog in the digital age by bringing together academic research and industrial requirements for discussing advantages and challenges of proposed solutions, with the goal to identify promising future directions.
Session 1: 14:00 - 15.15 Are Analog Electronics and Photonics Threatening the DSP?
Photonic accelerators based on recurrent optical spectrum slicing for 100 Gbaud and beyond
Adonis Bogris, University of West Attica
Dynamic Optical Processors: Shaping the Next Generation of Digital Signal Processing
Ana Gonzalez, Director of Strategic Partnerships at iPRONICS
100G/channel PAM4 Clock and Data Recovery using low-power analog SiGe circuits
Martin Bossard, Tetra Semiconductors Ltd.
High-speed analog neuromorphic computing for optical communication
Johannes Schemmel, University of Heidelberg
Panel discussion
Abstract Session 2
In the past decade, optical coherent transponders baudrate have grown from 28Gbaud to 140Gbaud. Still, the barriers to go at higher speed are high and the race toward higher baudrate may stop one day. The workshop will discuss the options for the future as well as the pros and cons. While most of transponders are “single wavelength”, some “Multi wavelengths” transponders have been introduced on the market by different industrial actors in the past, still, several of them decided to move back to a single wavelength transponder ASIC. Will the trend be reversed in the future? If yes, what would be the driving reasons? What benefit to expect from multi-wavelength transponder? Does it imply a single ASIC to modulate and demodulate the multiple wavelengths?
Session 2: 15:15 - 16:30 Single lambda at high baudrate or multi lambda transponder?
Stephen Adolph, Marvell
Kim Roberts, Ciena
Matthias Berger, Coherent
Pascal Pecci, Meta
Sebastian Randel, KIT
