Hybrid Ethernet PHY FEC Mode Selection for Latency and Burst Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Implementing high-speed Ethernet links using 100 Gb/s signaling poses challenges in managing and correcting errors, particularly due to the trade-off between latency and error propagation risks associated with different Forward Error Correction (FEC) modes.
Innovation Solution
A hybrid Physical Layer (PHY) approach is adopted, allowing for selective configuration of non-interleaved and interleaved RS-FEC modes during adaptive link training, enabling the determination of the appropriate FEC mode based on link conditions to request and confirm the use of either non-interleaved or interleaved FEC during link DATA mode through control and status fields in training frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If interleaved FEC mode is used, then error propagation risk is reduced, but latency increases
Solution Approach 1:
The system dynamically selects between interleaved and non-interleaved FEC modes based on real-time link conditions assessed during adaptive link training. The receiver evaluates channel characteristics and communicates feedback to the transmitter, enabling the FEC mode to adapt to varying link quality rather than being fixed, thus balancing error protection and latency requirements
2Loss of time
If non-interleaved FEC mode is used, then latency is reduced, but error propagation risk increases
Solution Approach 1:
The system dynamically selects between interleaved and non-interleaved FEC modes based on real-time link conditions. During adaptive link training, the receiver assesses channel characteristics and communicates feedback to the transmitter, enabling the FEC mode to adapt to varying link quality - using non-interleaved mode when latency is critical and link conditions permit, and switching to interleaved mode when error protection becomes more important
3Reliability
If adaptive link training with FEC mode determination is implemented, then link reliability is optimized, but protocol complexity increases
Solution Approach 1:
The system performs FEC mode determination during the link training phase, which occurs before normal data transmission begins. By establishing the optimal FEC mode in advance based on initial channel assessments, the system avoids the need for continuous mode switching or complex real-time decisions during data transmission, thereby optimizing reliability while limiting protocol complexity to the training phase only
Data Source
AI summary
Apparatus and methods for implementing high-speed Ethernet links using a hybrid PHY (Physical layer) selectively configurable to employ a non-interleaved RS-FEC (Reed Solomon Forward Error Correction) sublayer or an interleaved RS-FEC sublayer. An adaptive link training protocol is used during link training to determine whether to employ the non-interleaved or interleaved RS-FEC during link DATA mode. Training frames are exchanged between link partners including control and status fields used to respectfully request a non-interleaved or interleaved FEC mode and confirm the requested FEC mode is to be used during link DATA mode. The hybrid PHY includes interleaved RS-FEC and non-interleaved RS-FEC sublayers for transmitter and receiver operations. During link training, a determination is made to whether a local receiver is likely to see decision feedback equalizer (DFE) burst errors. If so, the interleaved FEC mode is selected; otherwise the non-interleaved FEC mode is selected or is the default FEC mode. The apparatus and methods may be implemented for 100GBASE-CR1 and 100GBASE-KR1 Ethernet links and interfaces.