SSD Polar Encoding with Dynamic Frozen Bit Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Polar codes used in solid state storage devices (SSDs) are non-universal, meaning their performance is significantly affected by channel signal-to-noise ratio (SNR), leading to suboptimal error correction capabilities as the Raw Bit Error Rate (RBER) changes over the device's lifetime.
Innovation Solution
A dynamic implementation of polar encoding and decoding is introduced, where the set of frozen bits is adjusted based on the current Raw Bit Error Rate (RBER) estimate, which is determined by monitoring parameters such as temperature, time since last writing, and number of program/erase cycles, allowing for adaptive error correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed set of frozen bits is used in polar code encoding, then the device complexity is reduced, but the error correction performance deteriorates as RBER changes over the device's lifetime
Solution Approach 1:
The patent applies dynamics by making the set of frozen bits adjustable rather than fixed. The controller dynamically selects from multiple pre-defined sets of frozen bits based on the current RBER estimate, allowing the error correction code to adapt to changing channel conditions over the device's lifetime while maintaining reasonable complexity through pre-computed sets.
Solution Approach 2:
The patent changes the parameter of frozen bit selection based on RBER estimates. By monitoring device aging parameters (temperature, time since last writing, program/erase cycles) and adjusting the frozen bit set accordingly, the system optimizes error correction performance for different operational conditions without requiring complex real-time code reconstruction.
2Reliability
If the set of frozen bits is adjusted dynamically based on RBER estimate, then the error correction performance is improved, but the device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-defining multiple sets of frozen bits during device manufacturing or initialization, each optimized for specific RBER ranges. This allows the controller to simply select from pre-computed sets based on current conditions, avoiding the need for complex real-time optimization calculations while still achieving adaptive error correction performance.
Solution Approach 2:
The patent implements feedback by continuously monitoring device aging parameters (temperature, time since last writing, program/erase cycles) to estimate current RBER, then using this feedback to select the appropriate frozen bit set. This closed-loop approach maintains optimal error correction performance without requiring complex real-time code design, as the feedback drives selection from pre-defined sets.
3Adaptability or versatility
If multiple sets of frozen bits are stored and selected based on RBER estimate, then the adaptability to varying operational conditions is improved, but the memory requirements increase
Solution Approach 1:
The patent applies segmentation by dividing the frozen bit configuration into multiple discrete sets, each optimized for specific RBER ranges. This segmentation allows the controller to store compact lookup tables or indices pointing to different frozen bit sets rather than storing complete code definitions, significantly reducing memory requirements while maintaining adaptability to varying operational conditions.
Data Source
AI summary
A method for operating a solid state storage device comprising memory cells exhibiting respective threshold voltage distributions comprises: providing sets of frozen bits each one associated with a respective RBER estimate being estimated according to a respective shape of the threshold voltage distributions; determining a current value of operative parameter(s) affecting the shape of the threshold voltage distributions; based on the current value of the operative parameter(s), determining a current shape of the threshold voltage distributions; determining a current RBER estimate associated with the current shape of the threshold voltage distributions; selecting a current set of frozen bits associated with the current RBER estimate; encoding the information bits and the current set of frozen bits with a polar code; storing the polar encoded bits in selected memory cells; reading the stored polar encoded bits, and decoding them according to said current set of frozen bits.