Storage Reliability Control for Faster Error-Tolerant Data Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage technologies, such as SSDs, employ extreme measures to ensure data accuracy but are costly and performance-intensive, and existing ECC protection schemes cannot dynamically adjust error correction levels based on application requirements, particularly for applications handling noisy data like images, sound, and video, which can tolerate errors.
Innovation Solution
Implementing dynamic reliability levels for non-volatile memory that allow host applications to specify varying degrees of data fidelity, enabling storage and retrieval with different levels of precision and accuracy, using error control coding and approximation levels to optimize performance without sacrificing application functionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If extreme error protection measures are implemented in storage devices, then data reliability is improved, but storage cost and performance overhead increase
Solution Approach 1:
The patent implements dynamic reliability levels that allow the storage system to adjust error correction intensity based on application requirements. The host can specify different reliability levels (e.g., high, medium, low) for different data sets, enabling the system to use stronger error correction only when necessary and weaker correction for tolerant applications, thus resolving the contradiction between reliability and cost/complexity
Solution Approach 2:
The system changes the parameter of error correction strength dynamically. By allowing host-specified reliability levels, the system can adjust the degree of error protection applied to data, matching the protection level to the actual needs of the application. This prevents over-protection of error-tolerant data while ensuring adequate protection for critical data
2Measurement precision
If strong error correction coding is used for all data, then data accuracy is improved, but read/write performance deteriorates
Solution Approach 1:
The patent enables dynamic adjustment of error correction strength based on host-specified reliability levels. Applications that prioritize performance over absolute accuracy (such as multimedia processing) can request lower reliability levels with corresponding performance benefits, while critical applications can request higher accuracy when needed, resolving the performance-accuracy tradeoff
Solution Approach 2:
The system applies partial error correction action by allowing hosts to specify minimum acceptable reliability levels. Instead of applying maximum error correction to all data, the system applies only the necessary level of correction for each data set, reducing overall processing overhead and improving performance for error-tolerant applications
3Reliability
If monitoring technology is deployed to detect device reliability indicators, then data protection capability is improved, but storage expense increases
Solution Approach 1:
The patent allows hosts to specify reliability levels as a parameter, enabling them to control the tradeoff between protection capability and expense. By making reliability levels adjustable and application-specific, the system prevents unnecessary expenditure on monitoring and protection for applications that can tolerate errors, while maintaining strong protection where needed
Data Source
AI summary
To address the storage needs of applications that work with noisy data (e.g. image, sound, video data), where errors can be tolerated to a certain extent and performance is more critical than data fidelity, dynamic reliability levels enable storage devices capable of storing and retrieving data with varying degrees of data fidelity to dynamically change the degree of data fidelity in response to an application's request specifying reliability level. By allowing the application to specify the reliability level at which its data is stored and retrieved, dynamic reliability levels can increase read/write performance without sacrificing application accuracy. The application can specify reliability levels for different types or units of data, such as different reliability levels for metadata as opposed to data and so forth.