Low-complexity symbol-rate equalization schemes for short-reach coherent-lite optical interconnects

Area of Science:

• Optical communications engineering
• Signal processing

Background:

• Short-reach optical interconnects are crucial for data centers.
• High computational complexity in traditional MIMO equalization limits performance.
• Coherent-lite systems require efficient equalization techniques.

Purpose of the Study:

• To propose low-complexity MIMO equalization schemes for short-reach optical interconnects.
• To reduce computational complexity in real-valued (RV) MIMO architectures.
• To evaluate the performance and robustness of proposed dynamic schemes.

Main Methods:

• Dynamic Pruning MIMO (DP-MIMO): Sparsifies filter coefficients, adjusting sparsity based on channel energy.
• Dynamic Clustering MIMO (DC-MIMO): Clusters filter coefficients with an energy-adaptive mechanism.
• Experimental validation in an 80-GBaud DP-16QAM system over 1-5 km SSMF links.

Main Results:

• DP-MIMO achieved ~40% reduction in real-valued multiplications (RMs) with a 0.20-dB ROP penalty.
• DC-MIMO achieved ~60% RMs reduction with a 0.15-dB ROP penalty.
• Both schemes demonstrated robustness against 2-ps IQ skew impairments.

Conclusions:

• DP-MIMO and DC-MIMO offer significant complexity reduction for coherent-lite optical interconnects.
• Dynamic adaptation to channel characteristics balances complexity and performance effectively.
• The proposed schemes are suitable for high-speed optical systems facing impairments like IQ skew.