Network Processor Packet Processing for CPU Load Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing edge router technologies rely solely on central processing units (CPUs) for service processing, which limits the efficiency due to the wide variety of services and the difficulty in improving forwarding and processing capabilities of the CPU.

Innovation Solution

Implementing a network processor (NP) in conjunction with the CPU to offload packet processing tasks. The NP receives packets, extracts features, looks up processing rules, and processes packets based on these rules, thereby reducing the CPU's processing load.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If all service processing is performed by the CPU, then the system maintains simplicity in architecture, but the service processing efficiency is low due to the wide variety of services and difficulty in improving CPU forwarding capabilities

Engineering Contradiction:
Improveservice processing efficiencyVSAvoidprocessing architecture complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent divides the packet processing function into two independent parts: the NP handles packet forwarding and basic processing tasks, while the CPU handles control plane functions and complex service processing. This segmentation allows each component to specialize in specific tasks, improving overall service processing efficiency while maintaining architectural clarity through defined functional boundaries.

Inventive Principle:
Principle #1Segmentation

2Speed

If the CPU processes all packets, then the processing logic is centralized and simple to manage, but the forwarding capability cannot be improved due to CPU performance limitations

Engineering Contradiction:
Improvepacket forwarding speedVSAvoidprocessing structure complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent extracts the packet forwarding function from the CPU and implements it in a dedicated NP. The NP handles time-critical packet forwarding operations at high speed, while the CPU focuses on control plane processing. This extraction resolves the contradiction by assigning speed-critical tasks to specialized hardware while maintaining management simplicity through centralized control plane processing.

Inventive Principle:
Principle #2Taking out (Extraction)

3Productivity

If multiple processors are introduced to improve processing capacity, then the processing power increases, but the system complexity increases and the CPU load for coordination increases

Engineering Contradiction:
Improveprocessing capacityVSAvoidmulti-processor coordination complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent assigns different processing qualities to different components: the NP is optimized for high-speed packet forwarding with simplified processing logic, while the CPU handles complex control plane functions. This local quality differentiation increases processing capacity for specific tasks without requiring complex multi-processor coordination, as each processor type is optimized for its designated function.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3731497B1Service processing method and network device
Publication Date: 2025.06.11 HUAWEI TECH CO LTD
  • EP3731497B1 patent drawingFigure 1~2
  • EP3731497B1 patent drawingFigure 3
  • EP3731497B1 patent drawingFigure 4

AI summary

This application provides a service processing method and a network device. The network device includes an NP and a CPU. The NP is configured to: receive a first packet; obtain a packet feature of the first packet; obtain a processing rule corresponding to the packet feature; process the first packet based on the processing rule, to obtain a second packet; and send the second packet to the CPU. The CPU is configured to: receive the second packet, and process the second packet. Before the CPU processes the first packet, the NP processes the first packet, and sends the second packet obtained after processing to the CPU, so that the CPU may directly process the second packet, to simplify a process in which the CPU processes the packet received by the CPU, thereby improving service processing efficiency of the network device.