Dynamic Candidate Flow Detection for Network Congestion

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Data networks face congestion due to 'elephant flows' that consume disproportionate network resources, leading to slow network response for all users, as they are designed for short bursts of data and not equipped to handle prolonged high usage.

Innovation Solution

Systems and methods to identify and manage candidate flows, such as elephant flows, by dynamically adjusting detection thresholds based on network environment, flow characteristics, and time, using techniques like video encoding rate analysis, buffer growth rate, and percentile calculations to prevent network degradation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If consumer access networks are designed for short bursts of data delivery, then the network can handle peak use demands of many users, but elephant flows that consume disproportionate network resources over extended periods cause network congestion and slow network response

Engineering Contradiction:

Improvenetwork throughputVSAvoidnetwork performance

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent applies dynamics by making the flow detection threshold adjustable rather than fixed. The threshold can be dynamically modified based on network conditions, flow characteristics, and management policies, allowing the network to adapt between detecting more flows (improving reliability) and detecting fewer flows (maintaining productivity) as needed

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of flow detection threshold from a static value to a dynamically adjustable one. By modifying this parameter based on various factors including flow type identification, buffer status, and network conditions, the system resolves the contradiction between maintaining high throughput and ensuring reliable performance

Inventive Principle:

Principle #35Parameter changes

2Reliability

If candidate flow detection threshold values are set low to identify more elephant flows, then more flows can be managed for optimal performance, but false positives increase and network resources are wasted on managing normal flows

Engineering Contradiction:

Improveflow identification accuracyVSAvoidnetwork resource consumption

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The patent applies local quality by making the detection threshold flow-specific rather than uniform across all flows. Different threshold values are applied to different flow types (e.g., video streaming vs. file downloads), allowing accurate identification of elephant flows in each category while avoiding false positives and unnecessary resource consumption

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system uses self-service by having flows effectively identify themselves as elephant flows through their own characteristics. By monitoring flow-specific metrics such as buffer growth rate, throughput, and duration, the system allows elephant flows to reveal their nature without requiring aggressive low-threshold detection, thus reducing false positives and resource waste

Inventive Principle:

Principle #25Self-service

3Loss of energy

If candidate flow detection threshold values are set high to reduce false positives, then fewer normal flows are misidentified, but elephant flows are missed and network congestion occurs

Engineering Contradiction:

Improvenetwork resource efficiencyVSAvoidnetwork throughput

Core Design Contradiction:

Loss of energyVSProductivity

Solution Approach 1:

The dynamic threshold adjustment allows the system to respond to changing network conditions. When congestion is detected or during peak periods, the threshold can be lowered to catch elephant flows that would otherwise be missed, thereby protecting network throughput while maintaining resource efficiency during normal conditions

Inventive Principle:

Principle #15Dynamics

4Reliability

If dynamic adjustment of candidate flow detection thresholds is implemented, then flow identification accuracy improves, but system complexity increases

Engineering Contradiction:

Improvecandidate flow detection accuracyVSAvoidthreshold management complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system reduces management complexity by making flows self-identifying through their inherent characteristics. By monitoring metrics that elephant flows naturally exhibit (high buffer growth, sustained throughput, long duration), the system automatically distinguishes them without requiring complex manual threshold management, thus maintaining high detection accuracy while limiting complexity growth

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS11677665B2Systems and methods for identifying candidate flows in data packet networks

Publication Date: 2023.06.13 OPANGA NETWORKS INC

AI summary

A computer-implemented method and a transport manager system operate to reduce network congestion by detecting one or more data flows in a network, determining, using a candidate flow detection threshold, whether a data flow of the one or more data flows is a candidate flow, the candidate flow detection threshold being based on one or more characteristics of the one or more data flows, and in response to determining that the data flow is the candidate flow, managing the data flow. A consumption rate, a duration, a number of bytes communicated, a throughput, or aggregated characteristics of the one or more data flows may be used to determine the candidate flow detection threshold.