Alvin Y. Sukmadji,Frank R. Kschischang
Abstract:Performance-complexity-latency trade-off curves for rate-0.88 concatenated outer Reed--Solomon codes and inner Chase-algorithm-based soft-decision Bose--Ray-Chaudhuri--Hocquenghem codes with PAM4 constellation using bit-interleaved coded modulation and multilevel coding coded modulation schemes over the AWGN channel are presented.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: in high - throughput communication systems (such as data center interconnects), how to find the optimal balance among performance, complexity, and latency. Specifically, the author studied the performance of the RS (Reed - Solomon) outer code and the soft - decision BCH (Bose - Chaudhuri - Hocquenghem) inner code concatenated coding scheme based on the Chase algorithm over the AWGN (Additive White Gaussian Noise) channel, and explored the performance - complexity - latency trade - offs of these coding schemes when using the PAM4 constellation.
### Main Problems
1. **Performance Optimization**: How to maximize the performance of the coding system while maintaining low complexity and low latency.
2. **Complexity Control**: How to reduce the computational complexity in the coding and decoding processes while ensuring performance.
3. **Latency Minimization**: How to reduce the processing latency of the coding system to meet the real - time communication requirements.
### Research Methods
To achieve the above goals, the author:
- Proposed a semi - analytical formula to determine the output frame error rate (FER) of the concatenated RS - SDBCH code decoded using the Chase - II algorithm.
- Evaluated the trade - offs in performance, complexity, and latency by computer - searching different combinations of inner and outer codes.
- Used Monte Carlo simulation to verify the accuracy of the proposed formula and carried out comparative analysis through experimental data.
### Key Formulas
The main formulas involved in the paper include:
1. **Generating Function**:
\[
W_{B,k,L}(x,y)=\left(1+(1 + x)^{B - 1}\right)^L(1 + x)^{k - BL}
\]
where the coefficients of \(W_{B,k,L}(x,y)\) represent the number of bit error patterns of length \(k\) bits, and these error patterns will lead to \(v\) RS symbol errors.
2. **Conditional Probability**:
- For the BICM scheme:
\[
\Pr(V_{i,j}=v)=\sum_{u = 0}^{k/2}\Pr(V_{i,j}=v|U_j = u)\Pr(U_j = u)
\]
where,
\[
\Pr(V_{i,j}=v|U_j = u)=\text{coeff}_{x^u y^v}\left(W_{B/2,k/2,L_{i,j}}(x,y)\right)\binom{k/2}{u}
\]
- For the MLC scheme:
\[
\Pr(V_{i,j}=v)=\sum_{u = 0}^{k}\Pr(V_{i,j}=v|U_j = u)\Pr(U_j = u)
\]
where,
\[
\Pr(V_{i,j}=v|U_j = u)=\text{coeff}_{x^u y^v}\left(W_{B/2,k,L_{i,j}}(x,y)\right)\binom{k}{u}
\]
### Conclusions
Through a large amount of code search and performance evaluation, the author found that:
- At the same complexity, some outer RS codes have better performance than KP4 codes.
- Compared with the BICM scheme, the MLC scheme can more effectively reduce the internal decoding operations at the same inner code rate, thereby improving the overall performance and reducing the complexity.
In summary, this paper aims to provide an effective coding scheme selection method for high - throughput communication systems to achieve the optimal balance of performance, complexity, and latency.