Intelligent Traffic Steering for Multi-Access Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems lack the ability for wireless communication devices to flexibly distribute traffic across 3GPP and non-3GPP access networks based on real-time link status and operator preferences, leading to inefficiencies and potential ping-pong effects in traffic steering, without considering overall network performance or operator policies.
Innovation Solution
Implement an intelligent traffic steering function within the wireless transmit/receive unit (WTRU) using machine learning components to dynamically manage uplink and downlink traffic across multiple access networks, incorporating operator preferences and network analytics to optimize traffic distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traffic steering decisions are made based on real-time link status and operator preferences using machine learning components, then network performance and resource utilization are improved, but device complexity increases due to the need for intelligent traffic steering function and analytics processing
Solution Approach 1:
The traffic steering function is segmented into multiple components: machine learning components for predictive analytics, intelligent decision-making modules, and traffic management units. This segmentation allows each component to handle specific aspects of traffic steering independently, reducing overall complexity while maintaining high network efficiency through specialized processing.
Solution Approach 2:
An intelligent traffic steering function acts as an intermediary between network analytics and traffic management. This intermediary layer processes real-time link status data, applies machine learning models, and generates steering decisions based on operator preferences, thereby improving network performance without directly increasing the complexity of individual network elements.
2Adaptability or versatility
If dynamic traffic distribution is implemented across multiple access networks, then adaptability to real-time conditions is improved, but signaling delays may increase due to continuous monitoring and decision-making processes
Solution Approach 1:
Machine learning components perform predictive analytics in advance to anticipate future network conditions and pre-determine optimal traffic steering decisions. This preliminary action allows the system to prepare traffic distribution strategies before actual network changes occur, improving adaptability while minimizing signaling delays by having decisions ready in advance.
Solution Approach 2:
The system implements feedback mechanisms that continuously monitor link status and network performance, comparing actual conditions against operator preferences and machine learning predictions. This feedback loop enables dynamic adjustment of traffic distribution while optimizing signaling efficiency by only triggering reconfiguration when significant changes are detected, rather than responding to every minor fluctuation.
3Measurement precision
If analytics-based operations are used for access traffic steering, then traffic steering accuracy is improved, but measurement and detection difficulty increases due to the need for real-time link status monitoring and analytics processing
Solution Approach 1:
The network elements perform self-monitoring and self-analytics by incorporating machine learning components directly within them. Each network element autonomously collects link status data, processes it through embedded analytics, and generates local steering decisions without requiring complex external monitoring systems. This self-service approach improves traffic steering accuracy while reducing the overall complexity of detection and measurement infrastructure.
Data Source
AI summary
Procedures, methods, architectures, apparatuses, systems, devices, and computer program products directed to analytics-based user plane optimization are provided. A first network entity (NE) may receive, from a second NE, information indicating analytics based operation for access traffic steering, splitting and switching (ATSSS), and first weights and a first weighting factors for multiple access networks; transmit, to a third NE, information indicating the analytics based operation, including the first weights and weighting factors; transmit, to a fourth NE, information indicating load metrics; receive, from the fourth NE, information indicating analytics for controlling traffic steering for a multi-access protocol data unit session, including second weights and/or second weighting factors for the multiple access networks, wherein the analytics are based on the load metrics and criteria for partitioning downlink traffic among the multiple access networks; and transmit, to the third NE, information indicating the second weights and/or second weighting factors.