Prefetch Controller Using Temporal Access Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity and computational intensity of tracking access patterns in storage systems make it challenging to determine data prefetch suitability in real-time, leading to inefficient performance improvements in disk and memory systems.
Innovation Solution
A storage control system with a prefetch controller that analyzes temporal data storage access patterns to determine data prefetch suitability, using a combination of software and hardware to identify suitable patterns for prefetching by monitoring and analyzing read/write operations, compressing data, and employing logical address mapping to reduce computational requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional prefetching methods are used with extensive access pattern tracking, then prefetch accuracy may improve, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent segments the access pattern analysis into distinct temporal components (recent accesses vs. historical patterns) and processes them separately through different data structures (LRFU stack vs. histograms), reducing the computational complexity of tracking while maintaining prefetch accuracy
Solution Approach 2:
The patent changes the parameter representation from tracking individual access sequences to using temporal histograms that count access frequencies within time windows, transforming the problem from sequence analysis to statistical analysis and thereby reducing computational requirements
2Measurement precision
If extensive access pattern tracking is performed to determine prefetch suitability, then prefetch prediction accuracy improves, but real-time determination becomes intractable
Solution Approach 1:
The patent performs preliminary organization of access patterns by maintaining an LRFU stack and temporal histograms during normal operation, so that when prefetch decisions are needed, the pre-organized data can be quickly queried without extensive real-time analysis
Solution Approach 2:
The patent replaces the mechanical process of tracking and analyzing individual access sequences with a statistical approach using temporal histograms that aggregate access patterns, enabling faster real-time determination while maintaining prediction accuracy
3Reliability
If the system keeps prefetched data cached for longer periods, then the data remains available for future accesses, but cache memory is wasted if the data is not accessed
Solution Approach 1:
The patent dynamically adjusts the prefetch keep time based on the temporal patterns observed in the histograms, setting the cache retention period to match the actual access intervals predicted from historical data, thereby optimizing both data availability and cache memory utilization
Data Source
AI summary
A storage control system includes a prefetch controller that identifies memory regions for prefetching according to temporal memory access patterns. The memory access patterns identify a number of sequential memory accesses within different time ranges and a highest number of memory accesses to the different memory regions within a predetermine time period.