Storage Access Pattern Analysis for Data Preloading
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems face challenges in accurately predicting which ranges in a storage area will be accessed next, leading to inefficient data preloading and resource allocation.
Innovation Solution
A method that calculates access numbers and variations for each range over multiple periods, identifies correlations between ranges, and uses clustering to group access patterns, facilitating the prediction of ranges likely to be accessed together.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If static patterns of data access are used for preloading, then processing performance is improved, but prediction accuracy of access ranges deteriorates
Solution Approach 1:
The storage area is divided into multiple ranges, and access patterns are analyzed separately for each range. This segmentation allows the system to capture localized access correlations that static global patterns miss, improving prediction accuracy while maintaining processing performance through targeted preloading of relevant ranges.
Solution Approach 2:
The system transitions from static access patterns to dynamic analysis by calculating access numbers and variations over multiple periods. This dynamic approach adapts to changing access behaviors, enabling accurate prediction of access ranges even as workloads evolve, thus resolving the contradiction between performance improvement and prediction accuracy.
2Ease of operation
If data is preloaded based on static patterns, then resource allocation is simplified, but prediction accuracy of access ranges deteriorates
Solution Approach 1:
The system automatically discovers access correlations by analyzing access numbers and variations across multiple periods without requiring manual configuration. This self-service approach maintains ease of operation while achieving high prediction accuracy, as the system adapts to access patterns autonomously and identifies correlated ranges for intelligent preloading.
3Device complexity
If access patterns are analyzed without considering variations over time, then analysis complexity is reduced, but prediction accuracy deteriorates
Solution Approach 1:
The system analyzes access patterns across multiple periods by calculating access numbers and their variations. This periodic analysis captures temporal dynamics in access behavior, improving prediction accuracy while managing complexity through structured multi-period comparison and correlation calculation between ranges.
Data Source
AI summary
A method performed by a computer for analyzing access to a storage device, the method includes: executing a calculating process that includes calculating, for each of a plurality of ranges obtained by dividing a storage area of the storage device, an access number or a variation of the access number in each of a plurality of periods, wherein the access number indicates the number of times of access to each of the plurality of ranges; and executing a determining process that includes determining a correlation between any two of the plurality of ranges in accordance with the access number or the variation of the access number for each period in each of the plurality of ranges.