Adaptive Flow Assignment for Heterogeneous Network Access

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing communication systems face challenges in efficiently routing traffic flows over multiple access networks with heterogeneous characteristics, such as varying available bit rates and delays, which affects Quality of Service (QoS) provisioning, especially in scenarios involving wireless links.

Innovation Solution

A method and communication system that utilize a Markov Decision Process (MDP) based control scheme to dynamically assign traffic flows across multiple access networks by monitoring and modeling network characteristics, optimizing flow allocation based on stochastic models of bit rate and delay, and periodically updating the policy for optimal network selection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If static flow assignment policies are used to route traffic flows over multiple access networks, then the network configuration is simple and stable, but the Quality of Service (QoS) provisioning is insufficient and packet delays increase

Engineering Contradiction:
ImproveQoS provisioningVSAvoidflow assignment control complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic flow assignment by continuously monitoring network characteristics (available bit rate, delay) and periodically reassigning traffic flows to optimal access networks based on real-time conditions. This dynamic approach replaces static assignment policies, enabling the system to adapt to changing network conditions and improve QoS provisioning while managing complexity through automated control loops.

Inventive Principle:
Principle #15Dynamics

2Loss of time

If dynamic flow assignment based on real-time network measurements is implemented, then QoS provisioning is improved and packet delays are reduced, but the system complexity and overhead increase

Engineering Contradiction:
Improvepacket delayVSAvoidnetwork monitoring and control complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The patent employs feedback mechanisms where the system continuously measures network characteristics (available bit rate, delay) and uses this information to adjust flow assignments. The feedback loop includes monitoring network conditions, comparing them against performance thresholds, and reassigning flows accordingly. This feedback-driven approach reduces packet delay by responding to real-time network changes while managing complexity through structured control algorithms.

Inventive Principle:
Principle #23Feedback

3Adaptability or versatility

If multiple access networks with heterogeneous characteristics are integrated, then network versatility and available bit rate increase, but network management complexity and heterogeneity challenges increase

Engineering Contradiction:
Improvemulti-network access capabilityVSAvoidnetwork management complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a universal flow assignment framework that can manage multiple heterogeneous access networks (wireless, wireline) using common monitoring and control mechanisms. The system assigns flows to any available access network based on standardized performance metrics (available bit rate, delay), enabling multi-network versatility while simplifying management through unified control policies that work across different network types.

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

Data Source

PatentEP1976318B1Method, system and device for optimal flow assignment for heterogeneous network access
Publication Date: 2019.05.15 DEUTSCHE TELEKOM AG
  • EP1976318B1 patent drawingFigure 1
  • EP1976318B1 patent drawingFigure 2
  • EP1976318B1 patent drawingFigure 3

AI summary

In a setting where a device (10) has access to multiple access networks (20, 30, 40), the distribution of traffic flows generated by applications running on a device amongst the access networks can enable better network utilization than single network use at a time. However, the variation in network characteristics like available bit rate and delay make the problem of flow assignment challenging. In the present invention we propose a method and system for an adaptive assignment of flows to different access networks, which results in a much better performance (in terms of packet losses, delays and allocated bit rates for the flows) than a static flow allocation policy. The adaptive flow reassignment is controlled by using stochastic characterization of the network characteristics of all access networks (20, 30, 40) and adopting a Markov Decision Process based selecting approach to optimally assign flows to the access networks (20, 30, 40).