SLC Cache Flush for Cold Data via Age Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory sub-systems using greedy garbage collection techniques inefficiently manage data blocks, leading to 'clogs' due to overlooking cold host data, resulting in reduced throughput and increased corruption risk, as they prioritize data blocks based solely on the amount of valid data rather than when the data was written.
Innovation Solution
Implementing a memory flush operation that switches the garbage collection priority from data amount to data age, relocating valid data from SLC cache to other memory types (TLC, MLC, QLC) based on a threshold condition related to the data's dwell time, ensuring efficient use of data blocks and reducing corruption by eliminating cold host data clogs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If greedy garbage collection prioritizes data blocks based solely on the amount of valid data, then the garbage collection process is simple to implement, but cold host data is overlooked causing clogs and reduced throughput
Solution Approach 1:
The patent changes the prioritization parameter from 'amount of valid data' to 'data age' (when data was written). This parameter change allows the system to identify and flush cold host data that has been sitting in the SLC cache for extended periods, preventing clogs and improving throughput while maintaining a relatively simple implementation approach.
Solution Approach 2:
The patent implements periodic flushing of the SLC cache based on data age thresholds. By periodically identifying and flushing old data rather than continuously managing all data blocks, the system simplifies the garbage collection process while effectively preventing clogs caused by cold host data accumulation.
2Productivity
If greedy garbage collection focuses only on valid data amount, then the data block management is efficient, but the risk of data corruption increases due to overlooked cold host data
Solution Approach 1:
The patent introduces data age as an additional parameter to the garbage collection prioritization criteria. By considering both valid data amount and data age, the system maintains efficient data block management while identifying cold host data that requires flushing, thereby reducing data corruption risk without sacrificing management efficiency.
3Quantity of substance
If the SLC cache retains cold host data for extended periods, then memory resource utilization is maximized, but energy consumption increases due to additional program-erase cycles
Solution Approach 1:
The patent implements periodic flushing of the SLC cache based on data age thresholds. By periodically ejecting cold host data before it becomes problematic, the system maintains high memory resource utilization while reducing the need for additional program-erase cycles, thereby lowering energy consumption.
Solution Approach 2:
The patent performs preliminary flushing of cold host data based on age criteria before it accumulates and causes problems. This preliminary action prevents the need for more intensive garbage collection operations later, reducing overall energy consumption while maintaining resource utilization.
4Device complexity
If greedy garbage collection only considers valid data amount, then the garbage collection algorithm is simple, but overwrite rates decrease due to clogs from cold host data
Solution Approach 1:
The patent changes the algorithm's prioritization parameter from solely 'valid data amount' to include 'data age'. This simple parameter addition enables the algorithm to identify cold host data for flushing, preventing clogs and improving overwrite rates while maintaining relatively low algorithmic complexity.
Data Source
AI summary
A method includes identifying a first block of a plurality of blocks stored at a first memory based on an amount of valid data of the first block, and writing the valid data of the first block from the first memory to a second memory. The first memory has a first memory type and the second memory has a second memory type different from the first memory type. The method further includes identifying a second block of the plurality of blocks stored at the first memory based on an age of valid data of the second block, determining that the age of the valid data of the second block satisfies a threshold condition, and in response to determining that the age of the valid data of the second block satisfies the threshold condition, writing the valid data of the second block from the first memory to the second memory.