Cache-Page Cgroup Tracking for Fair Buffered I/O Throttling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The accuracy of traffic statistics results for cgroups in Linux systems is low, leading to unfair and inaccurate throttling of buffered I/O operations.

Innovation Solution

A memory management method that updates a record table for each cache page to track which cgroups have read or written data, using a union set of cgroups and their access times to improve the accuracy of traffic statistics, enabling fairer throttling.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traffic statistics are collected for each cgroup in existing Linux systems, then throttling of buffered I/O can be performed, but the accuracy of traffic statistics results is low

Engineering Contradiction:
Improveaccuracy of traffic statisticsVSAvoidfairness of throttling
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent segments the traffic statistics by cache page, creating a record table for each cache page that tracks which cgroups have accessed the data and their respective access counts. This segmentation allows accurate attribution of traffic to specific cgroups at the cache page level, resolving the inaccuracy in existing system-wide or process-level statistics.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a feedback mechanism where the record table is updated after each read/write operation to reflect the current state of cgroup access statistics. This feedback loop ensures that throttling decisions are based on accurate, up-to-date statistics, improving both measurement precision and throttling fairness.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If a single record table is used for all cgroups, then device complexity is reduced, but measurement precision of individual cgroup traffic is lost

Engineering Contradiction:
Improvecgroup traffic statistics accuracyVSAvoidrecord table structure
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

Instead of using a single monolithic record table, the patent creates a separate record table for each cache page. Each record table is a small, simple structure that only needs to track cgroups accessing that specific cache page, reducing the complexity burden while enabling precise measurement.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent adds a new dimension of organization by structuring record tables at the cache page level rather than at the system level. This dimensional shift from global to local granularity enables precise cgroup traffic measurement without requiring a complex global tracking system.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS12399824B2Memory management method and apparatus
Publication Date: 2025.08.26 HUAWEI TECH CO LTD
  • US12399824B2 patent drawing
  • US12399824B2 patent drawing
  • US12399824B2 patent drawing

AI summary

This disclosure provides a memory management method and apparatus, to perform refined management on a process access status in a physical host, a virtual machine, or a container, and a traffic occupation status of each control group (cgroup). The method can includes a first record table corresponding to a target cache page is used to record all cgroups that have read or written target data from or into the target cache page and a quantity of times that the target data is read or written by all the cgroups, to improve accuracy of a traffic statistics result of a cgroup. In this case, throttling is performed on a first cgroup based on an updated first record table, so that throttling on the first cgroup can be more fair and accurate.