Flash Memory Garbage Collection Using Neural Network Valid Page Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional garbage collection operations in flash memory devices are inefficient, leading to decreased performance as the number and size of victim blocks increase, due to the time-consuming process of reallocating memory space.
Innovation Solution
A method and device that utilize a neural network model to estimate the valid page count of each memory block, selecting a victim block for garbage collection based on these estimates to optimize the garbage collection process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional garbage collection operations are performed by reading valid data from victim blocks and moving it to destination blocks, then free blocks are allocated for future writes, but the operation takes considerable time and efficiency decreases as the number and size of victim blocks increase
Solution Approach 1:
The patent applies preliminary action by estimating the valid page count of memory blocks in advance using a neural network model before selecting victim blocks for garbage collection. This allows the system to proactively identify suitable candidate blocks without having to read and analyze all blocks, thereby reducing the time required for garbage collection operations while still ensuring adequate free block allocation.
2Quantity of substance
If the number and size of victim blocks increase to allocate more free space, then more free blocks become available for writing, but the overall efficiency of garbage collection operations decreases
Solution Approach 1:
The patent applies self-service by using a neural network model that automatically learns and predicts the valid page count of memory blocks based on historical data and patterns. The model autonomously identifies suitable victim blocks without requiring manual intervention or exhaustive analysis, enabling the system to efficiently manage free block allocation even as the number of victim blocks increases, thereby maintaining garbage collection efficiency.
3Measurement precision
If exhaustive reading of all memory blocks is performed to accurately identify victim blocks, then precise selection of blocks for garbage collection is achieved, but the time and computational resources required increase significantly
Solution Approach 1:
The patent introduces an intermediary element - a neural network model - that acts as a mediator between the raw memory block data and the garbage collection decision-making process. The model processes and interprets block characteristics to predict valid page counts, providing accurate victim block identification without requiring exhaustive reading of all blocks, thus reducing the time and computational resources needed while maintaining precision.
Data Source
AI summary
Garbage collection is performed according to an estimated number of valid pages. A storage device estimates a valid page count at a future time based on a valid page count at each of past time steps and a present time step using a neural network model and selects a victim block that undergoes the garbage collection from memory blocks based on an estimated valid page count. A memory block having a lowest estimated valid page count or having an estimated valid page count having a maintaining tendency is selected as the victim block or a memory block having the estimated valid page count having a decreasing tendency is excluded from selecting the victim block.