Flash Storage Entropy-Based Data Routing for Endurance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flash storage devices in surveillance systems face inefficiency and reduced endurance due to constant bit rate (CBR) data storage, where less important data fills storage space indiscriminately, leading to increased bit-error-rate (BER) and potential data loss, especially when storing low activity scenes alongside high activity data.
Innovation Solution
A storage device controller identifies the entropy level of media streams and routes high entropy data to healthier blocks with lower BER and optimal write parameters, while storing low entropy data in blocks with higher BER and less stringent write parameters, thereby improving data protection and endurance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If CBR data storage is used in flash storage devices, then bandwidth management is simplified and encoding cost is reduced, but storage efficiency deteriorates and endurance decreases due to indiscriminate filling of storage blocks
Solution Approach 1:
The patent applies local quality by differentiating storage blocks into two types: first blocks with lower BER for high entropy data and second blocks with higher BER for low entropy data. This localized differentiation allows the system to optimize storage efficiency by placing less important low entropy data in higher BER blocks, thereby preserving endurance-critical blocks for more important data while maintaining CBR encoding simplicity.
Solution Approach 2:
The patent changes the parameter of block selection based on data entropy characteristics. By identifying entropy levels and dynamically routing data to appropriate block types, the system transforms the static storage approach into a dynamic one that adapts to data characteristics, improving overall storage efficiency without increasing encoding complexity.
2Ease of operation
If CBR data storage is used in flash storage devices, then bandwidth management is simplified, but storage endurance deteriorates due to indiscriminate filling of storage blocks
Solution Approach 1:
The patent preserves ease of bandwidth management by maintaining CBR encoding while applying local quality differentiation in the storage layer. By creating distinct block types with different BER characteristics and selectively placing data based on entropy, the system protects endurance-critical blocks from excessive wear caused by storing low-value data, thereby extending storage endurance without complicating bandwidth management.
Solution Approach 2:
The patent introduces an intermediary entropy identification mechanism that sits between the CBR encoding process and the physical storage blocks. This intermediary layer analyzes data characteristics and routes appropriate data to appropriate blocks, acting as a mediator that preserves the simplicity of CBR management while protecting storage endurance through intelligent data placement.
3Quantity of substance
If low entropy data is stored in the same blocks as high entropy data, then storage capacity is maximized, but data reliability deteriorates due to increased BER affecting important data
Solution Approach 1:
The patent applies segmentation by dividing storage blocks into two distinct categories: first blocks with lower BER for high entropy data and second blocks with higher BER for low entropy data. This segmentation allows the system to maintain overall storage capacity by utilizing both block types, while simultaneously protecting data reliability by ensuring that high entropy data is never placed in high BER blocks where errors would be more likely.
Solution Approach 2:
The patent implements local quality by assigning different quality levels (BER characteristics) to different storage blocks based on the importance of the data they will hold. High entropy data receives the quality protection of lower BER blocks, while low entropy data is intentionally placed in higher BER blocks where errors are less critical, thereby maintaining both capacity and reliability.
4Device complexity
If uniform write parameters are used for all data, then device complexity is minimized, but data protection deteriorates due to inability to optimize for different data importance
Solution Approach 1:
The patent applies local quality to write parameters by implementing different write latency and voltage settings for first blocks and second blocks. First blocks receive optimized write parameters with higher latency and appropriate voltage for maximum reliability, while second blocks use faster, less stringent parameters since they store less critical data. This localized parameter differentiation improves data protection without requiring complex dynamic adjustment mechanisms.
Solution Approach 2:
The patent introduces dynamics by making write parameters adaptive to data entropy and block type rather than uniform. The system dynamically selects write parameters based on the entropy identification result and target block type, allowing optimization of data protection for each data segment while keeping the overall control mechanism relatively simple through rule-based parameter selection.
Data Source
AI summary
Aspects of a storage device including a memory and a controller are provided which allow for data associated with a media stream and having high entropy to be stored in healthier memory locations, with improved data protection, and with more optimal NAND parameters than for data having low entropy. After receiving data associated with a media stream, the controller identifies an entropy level of the data. When the entropy level meets an entropy threshold, the controller stores the data in a first block of the memory associated with a lower BER, and/or with a higher write latency or a first, more discrete voltage. Alternatively, when the entropy level does not meet the entropy threshold, the controller stores the data in a second block of the memory associated with a higher BER, and/or with a lower write latency or a second, less discrete voltage.


