ECN Optimization Model for Adaptive Network Congestion Handling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Network congestion in data centers due to insufficient network device capacity or transient traffic bursts leads to high delays and degraded user experience, as network devices often discard packets, affecting real-time performance and application quality.

Innovation Solution

A network congestion handling method that uses an Explicit Congestion Notification (ECN) optimization model to determine reference probabilities for ECN configuration, allowing for adaptive ECN marking based on current network status, and updates the model using network analysis to improve trustworthiness and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If network device uses traditional packet discarding method during congestion, then device complexity is reduced, but service delay increases and user experience degrades

Engineering Contradiction:
Improvenetwork device complexityVSAvoidservice delay
Core Design Contradiction:
Device complexityVSLoss of time

Solution Approach 1:

The patent changes the parameter of packet handling from binary discard/forward to probabilistic forwarding based on ECN marking probability. The network device adjusts forwarding probability according to congestion conditions and ECN configuration, transforming a simple discard mechanism into an adaptive probabilistic one that reduces delay while managing complexity through standardized protocols.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If network device increases cache capacity to handle traffic bursts, then packet loss decreases, but device complexity and cost increase

Engineering Contradiction:
Improvepacket loss rateVSAvoidnetwork device complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces ECN marking as an intermediary mechanism between traffic congestion and packet handling decisions. Instead of relying solely on large caches to absorb bursts, the system uses ECN marks as a mediator signal that triggers probabilistic forwarding adjustments, reducing the need for oversized cache memory while maintaining reliability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent implements feedback through ECN configuration information that reflects network congestion status. The network device uses this feedback to dynamically adjust packet forwarding probability, creating a closed-loop system that adapts to traffic conditions without requiring increased cache capacity, thus maintaining reliability while controlling complexity.

Inventive Principle:
Principle #23Feedback

3Device complexity

If network device uses fixed ECN configuration, then device complexity is reduced, but adaptability to changing network conditions deteriorates

Engineering Contradiction:
ImproveECN configuration complexityVSAvoidadaptability to network conditions
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent transforms fixed ECN configuration into dynamic configuration by introducing probabilistic forwarding based on real-time congestion conditions. The network device adjusts ECN marking probability according to current network state, making the system adaptable to changing conditions while managing complexity through standardized dynamic adjustment mechanisms rather than multiple fixed configurations.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies preliminary action through pre-configured ECN parameters and probability ranges that the network device uses before actual congestion occurs. The system prepares adaptive forwarding strategies in advance based on predicted congestion patterns, enabling quick response to changing conditions without requiring complex real-time decision-making, thus balancing adaptability with complexity.

Inventive Principle:
Principle #10Preliminary action

4Device complexity

If network device discards packets during congestion, then device complexity is minimized, but information loss increases

Engineering Contradiction:
Improvepacket handling complexityVSAvoidinformation loss
Core Design Contradiction:
Device complexityVSLoss of information

Solution Approach 1:

The patent converts the harmful effect of congestion (which traditionally causes packet discard) into a beneficial signal for adaptive forwarding. ECN marks transform congestion information into a useful feedback mechanism that triggers probabilistic forwarding adjustments, turning what would be pure information loss into a controlled information-based decision process that minimizes unnecessary packet loss while maintaining simple handling logic.

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Data Source

PatentUS12107766B2Network congestion handling method, model update method, and related apparatus
Publication Date: 2024.10.01 HUAWEI TECH CO LTD
  • US12107766B2 patent drawing
  • US12107766B2 patent drawing
  • US12107766B2 patent drawing

AI summary

A network congestion handling method includes obtaining network status information; determining, based on the network status information and an Explicit Congestion Notification (ECN) optimization model, a reference probability corresponding to recommended ECN configuration information; determining destination ECN configuration information based on the reference probability; and performing ECN marking on a packet by using the destination ECN configuration information.