Marc Bohn, Coriant GmbH & Co. KG, Germany; Sebastian Randel, Karlsruhe Institute of Technology, Germany
Within the last decade, coherent DSP technology has emerged as the key enabler for optical transmission at rates from 100 Gbps up to 400 Gbps per wavelength. Today, around seven DSP solutions from different companies are offered, all competing to best answer to the operators needs such as performance, cost, and power.
Up to now, this competition seems to drive innovation in the direction of increased speed and capacity as vendors introduce high-performance soft-decision FEC codes, fiber nonlinearity compensation, and probabilistic constellation shaping. With all these advanced features, performance is getting closer and closer to the Shannon limit, making significant performance improvements in the range of >1dB unlikely to occur. At the same time, power consumption is getting more and more important and the timeline of new ASIC generations is following closer and closer the availability of new lower power CMOS process nodes, for which the end of Moore’s law has been predicted.
This brings up the question whether the industry as a whole would benefit from a successive standardization of coherent DSPs. Today, pretty much all coherent DSPs include a 100G DP-QPSK mode which is interoperable. However, it uses a hard-decision FEC which cannot compete with more advanced soft-decision FECs. Looking forward, the following questions arise:
What would it take to standardize higher-order modulation schemes e.g. 16QAM and 64-QAM as well as high-performance FECs?
Do operators see potential benefits in this?
Will standardization of coherent DSPs finally be driven by the need for high-capacity short-reach?
Is the optics market truly unique or will it ultimately be shared among 2-3 players (compare markets like CPU, GPU, LTE, PON, DSL, …)?
On this panel, we want to elude answers to these questions by bringing together speakers from key operators and system vendors.