Cache Replication Scoring Using IO Mode and Access Frequency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage systems fail to optimally replicate data blocks from a disk to a cache, as they only consider access frequencies without accounting for the IO modes, which affect the actual cache hit ratio.
Innovation Solution
A method that determines multiple access frequencies and IO modes of data blocks, calculates cache hit ratios corresponding to these IO modes, and assigns scores to data blocks based on their access frequencies and cache hit ratios to determine which data blocks should be replicated to the cache.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If data blocks are replicated to cache based only on access frequency, then the replication process is simple, but the cache hit ratio is suboptimal
Solution Approach 1:
The patent changes the parameters used for data block selection from solely access frequency to a composite metric incorporating both access frequency and IO mode characteristics. This allows the system to distinguish between different types of access patterns (sequential vs random, read vs write) and select data blocks that are more likely to yield actual cache hits, thereby improving the cache hit ratio without significantly increasing implementation complexity
Solution Approach 2:
The patent segments the data block selection process into distinct components: calculating access frequency, determining IO mode characteristics, computing prediction hit ratios, and generating final scores. This segmentation allows each component to be optimized independently and makes the overall system more maintainable while improving selection accuracy
2Reliability
If multiple factors (access frequency and IO mode) are considered for data block selection, then the cache hit ratio improves, but the calculation complexity increases
Solution Approach 1:
The patent introduces prediction hit ratio as an intermediary metric that synthesizes multiple factors (access frequency and IO mode characteristics) into a single interpretable score. This intermediary allows the system to consider multiple dimensions of data access behavior while maintaining a clean, manageable calculation framework that doesn't excessively increase complexity
Solution Approach 2:
The patent creates a composite scoring mechanism that combines access frequency and IO mode characteristics into a unified selection criterion. By weighting these different factors appropriately, the system achieves improved cache hit ratio while keeping the overall calculation structure relatively simple and manageable
3Ease of manufacture
If data blocks are selected for replication without considering IO mode, then the system is easier to implement, but the actual cache performance is suboptimal
Solution Approach 1:
The patent extends the selection criteria by adding IO mode characteristics as a new parameter dimension. This allows the system to distinguish between different access patterns and select data blocks that are more likely to be actually accessed, thereby improving cache performance while maintaining reasonable implementation ease through structured calculation approaches
Data Source
AI summary
Embodiments of the present disclosure relate to a method, a device, and a computer program product for storing data. The method includes determining multiple access frequencies and multiple input/output (IO) modes of multiple data blocks, where the multiple data blocks are stored in a disk, and an IO mode in the multiple IO modes indicates an access size, a read operation proportion, and a sequential access proportion. The method further includes determining multiple cache hit ratios corresponding to the multiple IO modes. The method further includes determining multiple scores of the multiple data blocks based on the multiple access frequencies and the multiple cache hit ratios. In addition, the method further includes determining, based on the multiple scores of the multiple data blocks, a data block in the multiple data blocks to be replicated to a cache.