Dynamic Processor Core Scheduling for Real-Time Network I/O

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current multiprocessor core systems face inefficiencies in processing network I/O requests due to static configuration methods, which fail to meet real-time demands and lead to suboptimal processor core scheduling, limiting overall system performance.

Innovation Solution

A dynamic method for scheduling processor cores that involves obtaining control parameters such as data flow indications, internal transmission durations, processor core usage, and application critical path information to dynamically switch packets between idle processor cores and adjust processing modes between interruption and polling modes based on overhead analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If static configuration is used for processor core scheduling, then system stability is maintained, but real-time network I/O processing demand cannot be met and system efficiency deteriorates

Engineering Contradiction:
Improvesystem stabilityVSAvoidsystem efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent implements dynamic processor core scheduling by monitoring system state parameters (CPU usage, memory usage, network I/O load) in real-time and automatically adjusting task-to-core bindings. The scheduler transitions from static configuration to dynamic reconfiguration based on current system conditions, allowing the system to adapt to varying workloads and meet real-time processing demands while maintaining stability through controlled transition mechanisms.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system dynamically changes scheduling parameters including processor core assignments, task priorities, and mode switchings (polling/interruption) based on monitored system state parameters. By adjusting these parameters in response to real-time conditions such as CPU utilization thresholds and network I/O intensity, the system optimizes processing efficiency without compromising stability.

Inventive Principle:
Principle #35Parameter changes

2Ease of operation

If tasks are bound to designated processor cores in static configuration, then scheduling simplicity is maintained, but processor core utilization becomes suboptimal and efficiency deteriorates

Engineering Contradiction:
Improvescheduling simplicityVSAvoidprocessor core utilization
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The scheduling system operates autonomously by automatically monitoring system state parameters and performing task-to-core rebindings without manual intervention. The scheduler evaluates current processor core utilization, task characteristics, and system load, then dynamically reassigns tasks to optimize core utilization while maintaining operational simplicity through automated decision-making algorithms.

Inventive Principle:
Principle #25Self-service

3Ease of manufacture

If processor core switching between interruption mode and polling mode is performed statically, then implementation simplicity is maintained, but processing efficiency for network I/O requests deteriorates

Engineering Contradiction:
Improveimplementation simplicityVSAvoidprocessing efficiency
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The system dynamically switches between interruption mode and polling mode based on real-time network I/O characteristics and system state parameters. The scheduler monitors packet arrival rates, processing delays, and CPU utilization to determine the optimal operating mode, transitioning dynamically to maximize processing efficiency for different network traffic patterns while maintaining implementation feasibility through standardized mode switching mechanisms.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP2701074B1Method, device, and system for performing scheduling in multi-processor core system
Publication Date: 2016.10.26 HUAWEI TECH CO LTD
  • EP2701074B1 patent drawingFigure 1
  • EP2701074B1 patent drawingFigure 2
  • EP2701074B1 patent drawingFigure 3a

AI summary

The present invention discloses a method, an apparatus, and a system for scheduling a processor core in a multiprocessor core system, which relate to the field of multiprocessor core systems, and can meet the demand for real-time network I/O processing, thereby improving the efficiency of the overall multiprocessor core system. The method for scheduling a processor core in a multiprocessor core system includes: obtaining, in the running process of the multiprocessor core system, a first control parameter, a second control parameter, a third control parameter, and a fourth control parameter; transferring a packet of a data flow that enters the multiprocessor core system to an idle processor core for processing based on the first control parameter, the second control parameter, and the third control parameter; and switching over the processor core in the multiprocessor core system between an interruption mode and a polling mode based on the fourth control parameter. The present invention is mainly applied to schedule a processor core.