Error Detection Device for Storage Apparatus Bit Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for detecting bit errors in storage apparatuses, such as semiconductor memories and hard disk drives, lead to increased storage space requirements due to the difference in write-data and read-data sizes, which hinders storage density improvement.
Innovation Solution
An error detection device and method that generates and uses error detection information associated with each piece of division-data to determine bit errors, allowing for efficient detection and correction of errors without increasing storage space by utilizing error detection codes that can identify error positions within the data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ECC code is added to detect bit errors in division-data with smaller read-data size, then error detection capability is improved, but storage density deteriorates
Solution Approach 1:
The patent divides the 64-bit data into two 32-bit division-data blocks, and calculates separate 6-bit ECC codes for each block. This segmentation allows the ECC code to be optimized for the smaller read-data size (32 bits) rather than the full write-data size (64 bits), reducing the overhead per block while maintaining error detection capability across the entire data structure.
Solution Approach 2:
The patent applies local quality by calculating ECC codes specifically tailored to each division-data block based on its actual size (32 bits), rather than using a uniform ECC code size for the entire 64-bit data. This allows each block to have the minimum necessary ECC overhead for its size, optimizing the balance between error detection and storage density.
2Measurement precision
If ECC code size is increased to accommodate different read-data sizes, then error detection accuracy is improved, but storage area increases
Solution Approach 1:
The patent changes the parameter of ECC code size from a fixed value to a variable value that depends on the read-data size. For 32-bit division-data, a 6-bit ECC code is used, which is the minimum necessary to detect and correct errors in that specific data size. This parameter adaptation ensures optimal error detection accuracy without unnecessary storage overhead.
3Reliability
If separate ECC codes are calculated for each division-data, then error detection capability for each block is improved, but processing complexity increases
Solution Approach 1:
The patent segments the error detection process by calculating separate ECC codes for each 32-bit division-data block. This segmentation enables independent error detection and correction for each block, improving reliability without requiring complex interaction between blocks. The segmented approach also simplifies the correction process by allowing targeted operations on specific blocks.
Solution Approach 2:
The patent applies partial action by calculating ECC codes only for the necessary 32-bit division-data blocks rather than processing the entire 64-bit data as a single unit. This partial approach reduces the processing complexity for each block while maintaining comprehensive error detection capability across the full data structure.
Data Source
AI summary
An error detection device includes a generation unit and a detection-correction unit. When writing data into a storage unit, the generation unit generates error detection information for each piece of division-data into which write-data are divided. The error detection information is information which is available in determining whether a bit error includes in the division-data, and by which a position where the bit error is caused can be detected based on plural pieces of the error detection information of the plural division-data composing the write-data. Using the error detection information, the detection-correction unit can determine whether the bit error is caused in the division-data. Using plural pieces of the error detection information of the data including the division-data in which it is detected that the bit error is caused, the detection-correction unit can detect the position where the error is caused, and can correct the bit error.