HM-SMR Disk Cache Partitioning by Write Update Frequency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in managing host managed-SMR (HM-SMR) disks is effectively cleaning the conventional magnetic recording (CMR) region to ensure available space, as write operations in SMR disks can destroy adjacent data, necessitating read and rewrite processes that impact performance.
Innovation Solution
A data processing method that classifies write data into different cache regions based on update frequency by determining a size threshold, dividing data blocks, and caching them in high and low frequency cache regions, using attribute information to manage metadata and traffic data blocks efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If data is cached in a single uniform cache region, then cache management is simple, but it cannot efficiently handle data with different update frequencies leading to suboptimal cache utilization
Solution Approach 1:
The cache region is segmented into multiple sub-regions (first cache sub-region and second cache sub-region) based on data update frequency characteristics. Hot data with high update frequency is stored in the first cache sub-region, while cold data with low update frequency is stored in the second cache sub-region. This segmentation allows the system to adapt to different data access patterns without requiring complex management mechanisms for each individual data block.
2Quantity of substance
If the CMR region is not cleaned timely, then available space is insufficient for new writes, but cleaning operations cause read-and-write overhead that degrades performance
Solution Approach 1:
The system performs preliminary classification of data blocks into hot and cold categories based on update frequency before cleaning operations. By identifying cold data blocks in advance that are unlikely to be accessed soon, the system can prioritize their eviction during cleaning operations. This preliminary action ensures that cleaning operations remove the least valuable data first, maintaining performance while ensuring adequate space availability.
Solution Approach 2:
The system implements a feedback mechanism that monitors data update frequency and dynamically adjusts cache management decisions. When the cache reaches capacity, the system uses observed update frequency patterns to determine which data blocks to evict. This feedback-driven approach ensures that cleaning operations are performed efficiently based on actual data access patterns, balancing space availability with performance preservation.
3Productivity
If write data is not classified by update frequency, then cache management is straightforward, but cache eviction policies cannot optimize for data access patterns reducing overall system efficiency
Solution Approach 1:
The system applies different quality characteristics to different parts of the cache based on local data properties. Specifically, it assigns different retention priorities to data blocks based on their update frequency characteristics. Hot data blocks receive higher retention priority and are kept in the cache longer, while cold data blocks receive lower priority and are evicted more readily. This local quality differentiation optimizes system efficiency by matching cache retention behavior to actual data access patterns without requiring complex global management policies.
Data Source
Figure 1~2
Figure 3~5
AI summary
The present disclosure provides a data processing method, a device, a computer apparatus (500) and a storage medium, wherein the method includes: in response to a target disk receiving at least one write request within a preset time period, determining a size threshold value for classifying a data update type according to a size of write data respectively indicated by each write request; determining a data update type corresponding to each write request according to a size of each write request and the size threshold value; dividing write data of the write request to obtain a data block according to a preset data block size, and caching the data block in a cache region of the target disk corresponding to the data update type, the target disk has multiple types of cache regions configured therein, different cache regions are configured to support caching of data with different update frequencies. Embodiments of the present disclosure divide data blocks of different data update types into different cache regions, which can efficiently manage write data of different update frequencies.