Cache Resource Eviction Using Time Heat Factors for Large Disks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cache eviction algorithms, particularly the greedy dual size frequency (GDSF) method, provide inaccurate resource heat calculations for resources stored in large disks, leading to reduced eviction accuracy.
Innovation Solution
A resource eviction method that calculates resource heat by acquiring access day numbers and frequencies, using time heat factors and update values to improve accuracy, and evicts resources based on these metrics, optimizing for large disks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the cache eviction algorithm based on GDSF is used to compute resource heat, then the resource heat can be obtained accurately for resources located in a small disk, but the resource heat has low accuracy for resources located in a large disk
Solution Approach 1:
The patent applies local quality by treating small disks and large disks differently through disk-specific parameters. For small disks, the original GDSF algorithm parameters are used, while for large disks, adjusted parameters (such as modified time heat factor calculation and access frequency thresholds) are applied to account for the different access patterns and resource characteristics, thereby achieving accurate resource heat computation for both disk types
Solution Approach 2:
The patent changes key parameters of the GDSF algorithm based on disk size. Specifically, it modifies the time heat factor calculation parameters and access frequency thresholds to adapt to large disk characteristics. These parameter adjustments allow the algorithm to maintain accuracy when computing resource heat for resources in large disks, while preserving the original accurate performance for small disks
2Productivity
If the resource heat computation uses simple access frequency metrics, then the computation speed is fast, but the eviction accuracy is reduced for large disks
Solution Approach 1:
The patent introduces dynamic parameter adjustment based on disk size. The algorithm dynamically selects different computation strategies: for small disks, it uses the original fast GDSF computation, while for large disks, it activates enhanced computation modes with adjusted parameters that prioritize accuracy. This dynamic adaptation allows the system to maintain both speed and accuracy by choosing the appropriate computation depth based on the specific disk context
Data Source
AI summary
A resource eviction method, an electronic device and a readable storage medium, which relate to the field of artificial intelligence technologies, such as cloud service technologies, big data technologies, or the like, are disclosed. The resource eviction method includes: acquiring an access day number of at least one target resource in a current cache period, and acquiring an access frequency of the at least one target resource according to the access day number and a preset time interval; acquiring a time heat factor corresponding to the current cache period, and acquiring resource heat of the at least one target resource according to the access frequency and the time heat factor; acquiring target heat according to the time heat factor; and evicting the target resource with the resource heat smaller than or equal to the target heat, updating the time heat factor according to a preset update value, and taking the updated time heat factor as a time heat factor corresponding to a next cache period.