Memory Data Migration by Periodic Sampling of Page Access Heat
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory data migration methods based on access frequency statistics are inefficient due to long cycles, leading to incorrect storage of frequently or infrequently accessed data, affecting system performance and causing delays.
Innovation Solution
A method for migrating memory data that involves sampling access requests in a periodic sampling mode, collecting statistics on access frequencies, and triggering data migration between different memory types based on determined access temperatures to ensure accurate identification of frequently and infrequently accessed data, facilitating migration to near and far memories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If access frequencies are collected based on long statistical cycles, then measurement accuracy of access frequencies is improved, but data processing speed deteriorates and system delay increases
Solution Approach 1:
The patent implements periodic sampling of access requests at predetermined intervals rather than continuous monitoring. The sampling module periodically collects access request information within sampling windows, enabling the system to balance statistical accuracy with timely data processing by updating access frequency statistics at regular intervals rather than waiting for long statistical cycles to complete
Solution Approach 2:
The patent performs preliminary sampling and statistical collection of access frequencies before actual data migration decisions are made. By pre-collecting and pre-processing access request statistics during sampling periods, the system prepares migration candidates in advance, reducing the overall time required for data migration operations and improving response speed
2Measurement precision
If access frequencies are collected based on long statistical cycles, then measurement accuracy of access frequencies is improved, but time consumption increases
Solution Approach 1:
The system uses periodic sampling with configurable sampling intervals and windows to collect access frequency statistics. This approach allows the system to gather sufficient statistical data for accurate measurements while limiting the time spent on statistics collection through defined sampling periods, preventing excessive time consumption
Solution Approach 2:
The patent samples a subset of access requests within sampling windows rather than analyzing every single access request in real-time. This partial sampling approach provides sufficiently accurate statistical information for data migration decisions while significantly reducing the time required for complete statistics collection
3Productivity
If data migration is based on inaccurate access frequency statistics, then data processing speed may be maintained, but system performance deteriorates due to incorrect storage allocation
Solution Approach 1:
The system continuously monitors access request patterns and uses this feedback to dynamically adjust data migration decisions. The sampling module collects ongoing statistics about which data blocks are frequently accessed, and this feedback loop ensures that hot data is consistently identified and migrated to high-speed storage, maintaining both speed and allocation accuracy
Solution Approach 2:
The patent implements dynamic data migration based on changing access patterns. Rather than static allocation, the system continuously updates access frequency statistics and adapts migration decisions accordingly, ensuring that storage allocation remains optimal even as data access patterns evolve over time
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for migrating memory data, a related apparatus, and a computer device are disclosed, and relate to the computer field. The method includes: sampling a plurality of access requests in a periodic sampling mode indicated by a sampling rule; collecting statistics on access frequencies of pages indicated by a sampling result, to determine access temperatures of the pages based on the access frequencies of the pages; and triggering data migration based on the access temperatures. The access requests are sampled in a plurality of manners such as the periodic sampling mode according to the sampling rule, to ensure statistical accuracy of the access frequencies of the pages, and reduce an amount of statistical data of the access frequencies of the pages, so that a processor can obtain the access frequencies of the pages as quickly as possible. When data migration is triggered based on the access temperatures of the pages, hot data may be migrated to a near memory, and cold data may be migrated to a far memory, so that the processor can obtain frequently accessed data from the near memory as quickly as possible. This improves a data processing speed of the system and reduces a data processing delay.