Upstream FEC Codeword Filling for Variable Burst Sizes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Passive Optical Networks (PONs), especially Ethernet Passive Optical Networks (EPONs), the challenge lies in efficiently encoding and decoding upstream FEC codewords without prior knowledge of data burst size, as the data burst length is not communicated from the MAC layer to the PHY layer, leading to inefficiencies in aggregate code rate and issues with sparse last shortened FEC codeword decoding.
Innovation Solution
The implementation of a codeword filling system that selects, combines, and shortens multiple FEC codeword sizes and rates, allowing for reversible upstream FEC encoding and decoding of variable-sized data bursts by using mixed codeword sizes, including shortened short, medium, and long codewords, to maximize aggregate code rate and efficiently handle variable data burst sizes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed codeword size is used for upstream FEC encoding, then the encoding process is simplified, but the aggregate code rate decreases when data burst size varies
Solution Approach 1:
The patent implements dynamic codeword size selection where the FEC encoder adapts the codeword size (short, medium, or long) based on the actual data burst length. This allows the system to optimize the aggregate code rate for each transmission by matching the codeword size to the data burst size, rather than using a fixed size that would result in inefficiency for variable-length bursts.
2Productivity
If data burst size information is communicated from MAC layer to PHY layer, then optimal codeword size selection is enabled, but the system complexity and overhead increase
Solution Approach 1:
The patent enables the PHY layer to autonomously determine the appropriate codeword size by analyzing the data burst length directly available at the PHY layer, without requiring explicit information exchange with the MAC layer. The PHY layer self-services the codeword size selection function using locally available information about the data burst.
3Productivity
If variable-sized codewords are used to match data burst size, then aggregate code rate is maximized, but the decoding complexity increases
Solution Approach 1:
The patent segments the variable-sized data burst into multiple codewords of different sizes (short, medium, long) based on the burst length. The decoding process similarly segments the received signal into corresponding codeword sizes for parallel decoding. This segmentation approach manages decoding complexity by processing different segments independently rather than requiring a single complex variable-size decoder.
4Productivity
If shortened FEC codewords are used for small data bursts, then transmission efficiency is improved, but decoding reliability decreases due to sparsity
Solution Approach 1:
The patent changes the codeword size parameter dynamically based on the data burst size. For small data bursts, it selects short codewords with appropriate parameters optimized for that size, rather than using fixed-length codewords with poor parameters for small bursts. This parameter adaptation ensures both efficiency and reliability by matching the codeword parameters to the actual data size being transmitted.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A device for codeword filling comprises at least one processor circuit. The at least one processor circuit is configured to receive portions of a data burst, encode the portions into blocks, and add the blocks to a buffer. The at least one processor circuit is configured to generate a first codeword from blocks of the buffer when a number of blocks in the buffer satisfies a threshold, remove the blocks from the buffer, and provide the first codeword for transmission. The at least one processor circuit is configured to generate a set of codewords from remaining blocks of the buffer when a marker indicating a data burst end is detected, the set of codewords being determined based at least on a number of the remaining blocks in the buffer when the marker is detected. The at least one processor circuit is configured to provide the set of codewords for transmission.