Packet Forwarding Burst Sequence Handling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current packet forwarding devices face challenges in improving packet processing capability and reducing forwarding delay due to cache pressure and poor bandwidth adaptability, especially when handling multiple service flows and cross-connected rigid pipes.

Innovation Solution

The method involves determining whether a target packet is a header or following packet in a burst sequence, using a feature field as a burst identifier to obtain port information from a stored correspondence, and sending the packet through the corresponding egress port, thereby reducing the need for complex table lookups and increasing forwarding rate and reducing delay.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the transmit end device caches multiple consecutive packets to form a burst data unit, then the forwarding rate is improved by reducing table lookup frequency, but the cache pressure on the packet forwarding device increases

Engineering Contradiction:
Improveforwarding rateVSAvoidcache pressure
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The patent segments packets into header packets and following packets within a burst sequence. Header packets trigger table lookups and egress port determination, while following packets are forwarded directly using pre-determined egress ports. This segmentation reduces the frequency of table lookups and egress port determinations, improving forwarding rate while minimizing cache requirements compared to buffering entire burst data units.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs egress port determination when a header packet is received, before processing following packets in the burst sequence. The egress port information is cached and reused for all subsequent following packets until the next header packet arrives. This preliminary action eliminates the need to cache multiple complete packets, reducing cache pressure while maintaining high forwarding rates.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If the packet forwarding device performs table lookup and forwarding for all packets one by one, then the forwarding accuracy is maintained, but the forwarding delay increases due to queuing and uncertain processing time

Engineering Contradiction:
Improveforwarding accuracyVSAvoidforwarding delay
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent merges the processing of multiple following packets by determining the egress port once for the entire burst sequence when the header packet is received. All following packets in the sequence share the same egress port determination, eliminating repeated table lookups and reducing queuing delay while maintaining accurate forwarding through the pre-determined egress port.

Inventive Principle:
Principle #5Merging (Combining)

3Ease of manufacture

If bandwidth of a physical port is divided into multiple timeslots with fixed bandwidth, then the network management becomes simpler, but the adaptability to service flow bandwidth requirements deteriorates

Engineering Contradiction:
Improvenetwork management simplicityVSAvoidbandwidth adaptability
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic burst sequence identification using feature fields in packet headers, allowing the system to adapt to varying service flow requirements in real-time. Unlike fixed timeslot allocations, the feature field-based approach dynamically identifies and processes burst sequences based on actual traffic patterns, providing bandwidth adaptability while maintaining simple forwarding operations through direct egress port determination.

Inventive Principle:
Principle #15Dynamics

4Reliability

If one service flow is corresponding to one rigid pipe, then the service quality is guaranteed, but the device complexity increases when handling a large quantity of cross-connected rigid pipes

Engineering Contradiction:
Improveservice qualityVSAvoidcross-connection complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent creates a universal forwarding mechanism that handles multiple service flows through a common feature field-based identification system. Instead of requiring separate rigid pipe infrastructure for each service flow, the system uses feature fields to universally identify burst sequences and determine egress ports, reducing cross-connection complexity while maintaining service quality guarantees through consistent forwarding rules.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3672167B1Packet forwarding method and device
Publication Date: 2023.08.02 HUAWEI TECH CO LTD
  • EP3672167B1 patent drawingFigure 1~2
  • EP3672167B1 patent drawingFigure 3A
  • EP3672167B1 patent drawingFigure 3B-1

AI summary

This application discloses a packet forwarding method and apparatus, and pertains to the field of communications technologies. The method includes: determining whether a current to-be-sent target packet is a header packet or a following packet in a burst sequence; when the target packet is the following packet in the burst sequence, obtaining port information of an egress port used to send the burst sequence, where the port information of the egress port is determined and stored when the header packet in the burst sequence is sent, and the burst sequence is a sequence including a plurality of packets that belong to a same service flow; and sending the target packet through the obtained egress port. Because the port information of the egress port of the burst sequence is determined and stored when the header packet in the burst sequence is sent, in this application, when sending the target packet, the target packet may be directly sent through the corresponding egress port determined based on the stored egress port information, and there is no need to perform a complex action of looking up a routing table or looking up another forwarding table again based on forwarding information carried in the target packet, thereby increasing a forwarding rate and reducing a forwarding delay.