Zoned Storage Garbage Collection Using Aggregated Valid Data Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Garbage collection in zoned storage devices (ZNS) affects normal reading and writing performance due to increased query requests and data migration pressure as the number of data entries in zones grows, leading to reduced input/output per second (IOPS) and bandwidth.
Innovation Solution
Determine valid data segments in batches using file object metadata, aggregate them into cache pools based on file objects, and write them into target zones sequentially to reduce metadata queries and IOPS, ensuring continuous data segments for efficient storage space collection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If garbage collection is performed in ZNS device by migrating valid data entry by entry, then storage space can be collected, but metadata queries increase and reading/writing performance deteriorates
Solution Approach 1:
The patent merges multiple valid data entries into a single aggregated data segment before migration. Instead of processing each data entry individually during garbage collection, the method combines adjacent valid data entries into one continuous data segment, reducing the number of metadata queries and IOPS operations required, thereby maintaining reading and writing performance while still achieving storage space collection.
Solution Approach 2:
The patent performs preliminary aggregation of valid data entries into continuous segments before the actual migration process. By pre-processing the data to create aggregated segments that can be migrated as single units, the system reduces the overhead of individual entry processing during garbage collection, thus preserving performance while enabling effective storage space reclamation.
2Ease of manufacture
If data entries are stored in zones, then sequential writing is enabled, but garbage collection impact increases with more data entries
Solution Approach 1:
The patent combines multiple data entries into aggregated segments during garbage collection, reducing the number of operations required to migrate valid data. This merging approach maintains the sequential writing capability of ZNS devices while minimizing the performance impact of garbage collection by reducing metadata queries and IOPS operations.
3Reliability
If valid data is migrated entry by entry, then data can be preserved, but IOPS and bandwidth are reduced
Solution Approach 1:
The patent merges adjacent valid data entries into single aggregated segments before migration. This approach preserves all valid data reliably while significantly reducing the number of IOPS and bandwidth operations required during garbage collection, as aggregated segments are migrated as single units rather than individual entries.
Solution Approach 2:
The patent performs preliminary aggregation of valid data entries into continuous segments before the migration process begins. This pre-processing step ensures data preservation while optimizing the subsequent migration operation to require fewer IOPS and bandwidth operations, thus maintaining system productivity.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Provided in examples of the disclosure are a method and device for garbage collection of a zoned storage device, and a storage medium. Valid data segments in a source zone to be collected in the zoned storage device are determined (S201); each valid data segment is read (S202) from the source zone, the read valid data segments are written into cache pools, and the valid data segments in the same cache pool are aggregated into a valid aggregated data segment; and a target cache pool to be written into the zoned storage device is determined (S203), the valid aggregated data segment in the target cache pool is written into a target zone, and storage space of the source zone is collected (S204).