Multicore Packet Processing via Dynamic Core Allocation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Multicore systems in mobile devices face inefficiencies in high-rate network data processing, leading to degraded traffic processing performance and increased battery consumption due to inefficient packet processing, particularly when handling high-volume traffic such as online games or real-time video streaming.

Innovation Solution

A method is introduced where selected cores in a multicore system are dynamically allocated to process packets based on data reception speed, context information, and packet throughput, allowing for parallel processing and efficient distribution of packets across multiple cores to optimize network layer and transport layer processing, thereby reducing power consumption and enhancing processing performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multicore systems are used for high-rate network data processing, then processing capacity increases, but packet processing efficiency deteriorates due to improper distribution

Engineering Contradiction:
Improveprocessing capacityVSAvoidpacket processing efficiency
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The patent segments packet processing by creating separate processing queues for different packet types (first packet flow and second packet flow) and assigning different cores to process each queue. This segmentation allows parallel processing while maintaining efficiency by preventing context switching between different packet types on the same core.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by dedicating specific CPU cores to specific packet processing queues based on their characteristics. High-priority packets are assigned to cores optimized for low-latency processing, while other packets use different cores, creating localized optimization for different traffic types.

Inventive Principle:
Principle #3Local quality

2Productivity

If multiple cores process packets in parallel, then throughput increases, but context switching overhead increases

Engineering Contradiction:
ImprovethroughputVSAvoidcontext switching overhead
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent segments packet flows into distinct queues (first packet flow queue, second packet flow queue) that are processed by different cores simultaneously. This segmentation eliminates context switching between packet types on the same core while maintaining high throughput through parallel processing across multiple cores.

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If packets are concentrated on one core for processing, then processing simplicity increases, but traffic processing performance deteriorates

Engineering Contradiction:
Improveprocessing simplicityVSAvoidtraffic processing performance
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent segments the packet processing load by creating multiple queues and distributing them across multiple cores. This maintains simplicity within each core's processing context while achieving high traffic processing performance through parallel execution on multiple cores.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from single-core sequential processing to multi-core parallel processing by adding the dimension of parallel execution. Multiple cores process different packet queues simultaneously, dramatically improving traffic processing performance while maintaining simplicity through clear queue-core mapping.

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

Data Source

PatentEP3404533B1Method and apparatus for data processing based on multicore
Publication Date: 2021.05.05 SAMSUNG ELECTRONICS CO LTD
  • EP3404533B1 patent drawingFigure 1A
  • EP3404533B1 patent drawingFigure 1B
  • EP3404533B1 patent drawingFigure 2

AI summary

A communication technique for combining a fifth generation (5G) communication system that supports higher data transmission rates after fourth generation (4G) systems with Internet of things (IoT) technology and to the system therefor are provided. The disclosure relates to a multiple cores-based data processing method and apparatus, the method including determining whether a condition for a first packet flow among one or more packet flows for data communication is met, if the condition is met, selecting cores to distribute packets of the first packet flow from among the multiple cores, allocating packets received through the first packet flow to the selected cores, and operating the selected cores in parallel to process the packets allocated from the first packet flow.