Are We Done With Digital Processing Innovation And Is Integration All That’s Left To Do?
Tuesday, 5 March, 19:30 - 21:30
Organizers: David Plant, McGill University, Canada; Peter Winzer, Nokia Bell Labs, USA
Provocateurs: To be determined.
Description: Progress in optical fiber communications up until ~2000 was mostly due to advances in device engineering, enabling more stable lasers, more narrow-band optical filters, more broadband optical amplifiers, lower-loss and lower-nonlinearity fibers forming dispersion-managed links, and higher-speed transponders, mostly based on simple on/off keying. Since then, progress has been mostly fueled by advances in digital communication techniques, including modulation, coding, and coherent digital signal processing, bringing systems close to their fundamental limits. Spatial parallelism, coupled with tight integration, seems to be the next (and maybe the last?) avenue to increase capacity. Will integration alone suffice to reduce cost and energy per bit at a rate our industry has gotten accustomed to? Are we done with research into digital processing techniques and is device integration all that is left to advance optical fiber communications? Are we missing techniques that can continue to scale up transmission capacities? Will Machine Learning save the day? On the other hand, short-reach transmission links are still far away off any fundamental limit. How can novel transmission technologies help reduce the cost and energy consumption of shorter-reach links in an application-relevant way?
Questions for Discussion:
- Short introductory presentation by session organizer
- One slide presentations from diverse group of industry provocateurs
- Vigorous audience participation after each presentation, with organizer facilitating wide ranging discussion
- Attendees come prepared with tough questions and insightful comments, and challenge the presenters