Semiconductor Memory Discarding Unit for SSD Cache Performance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
NAND memory-based SSDs face high processing loads due to compaction and garbage collection processes, which hinder high-speed implementation as cache for HDDs.
Innovation Solution
A semiconductor memory device with a discarding unit that selectively discards valid data based on high compaction cost conditions, a compaction unit that writes valid data to new blocks, and a controller that responds to read requests for discarded data as unreadable, thereby reducing processing loads and securing writable areas without executing compaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If compaction and garbage collection processes are performed in NAND memory, then data consistency and storage management are maintained, but processing load increases and high-speed implementation is prevented
Solution Approach 1:
The patent extracts the compaction process by introducing a discarding unit that selectively discards valid data based on determination information, separating the data discarding function from the traditional compaction process. This allows the compaction unit to operate more efficiently with reduced processing load while maintaining data consistency through controlled discarding operations.
Solution Approach 2:
The patent segments the data management functions into distinct units: a discarding unit that determines which data to discard based on determination information, a compaction unit that handles the actual compaction process, and a controller that coordinates operations. This segmentation allows each unit to specialize in specific tasks, improving overall processing efficiency while maintaining reliability.
2Area of stationary object
If compaction process is executed to manage valid data, then writable area is secured, but processing load increases
Solution Approach 1:
The patent applies preliminary action by having the discarding unit determine and discard valid data in advance based on determination information before the compaction process is fully executed. This preliminary discarding of unnecessary data reduces the amount of data that needs to be processed during compaction, thereby reducing processing load while still securing adequate writable area.
Solution Approach 2:
The patent applies local quality by selectively discarding only the valid data that meets specific criteria (based on determination information) rather than discarding all data uniformly. This localized approach to data discarding ensures that only unnecessary data is removed, preserving writable area while minimizing processing load.
3Productivity
If valid data is discarded without compaction, then processing load is reduced and high-speed performance is achieved, but data consistency may be compromised
Solution Approach 1:
The patent implements feedback mechanisms where the discarding unit continuously monitors determination information about valid data and adjusts discarding operations accordingly. The controller receives feedback from the discarding unit and compaction unit, coordinating operations to ensure data consistency is maintained while achieving high-speed performance through efficient data discarding.
Solution Approach 2:
The discarding unit operates autonomously by determining which valid data to discard based on determination information without requiring constant external intervention. This self-service capability allows the system to maintain data consistency through intelligent, automated discarding decisions while achieving high processing speeds.
Data Source
AI summary
According an embodiment, a semiconductor memory device includes a semiconductor memory chip to store plural pieces of data that are written and read in units of a page and are erased in units of a block including plural pages; a discarding unit to discard, after the data is written in the semiconductor memory chip with a logic address being designated, at least a portion of valid data among the plural pieces of data; a compaction unit to write the valid data excluding the discarded data in a second block among the valid data stored in a first block and erase the first block; and a controller to output, in response to a request for reading the discarded data, a response indicating that the data is unable to be read. When all the valid data included in a block are discarded, the discarding unit erases the block.