SDN Traffic Steering Controller for Network Congestion Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network management systems, such as those using ANDSF, lack dynamic congestion control and fail to anticipate future network states, leading to inefficiencies in traffic distribution across cellular and WiFi networks, especially in densely populated areas with ultra-dense cellular networks.
Innovation Solution
Integration of wireless user terminals with multihoming capabilities into Software Defined Network (SDN) architectures, allowing SDN controllers to issue access network selection commands based on terminal-side information and predictions, thereby optimizing network traffic through a traffic steering system that includes a traffic steering controller, terminal traffic steering agent, user profiling manager, network metrics manager, and access discovery and selection manager.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If static ANDSF rules are used for access network selection, then the terminal can discover and select alternative access networks, but the network cannot dynamically adapt to changing traffic conditions and congestion states
Solution Approach 1:
The patent implements dynamic traffic steering rules that are continuously updated based on real-time network conditions, terminal behavior patterns, and congestion states. The system transitions from static ANDSF rules to dynamic, adaptive routing decisions that respond to changing network environments, thereby improving network adaptability while managing complexity through automated learning mechanisms.
Solution Approach 2:
The system employs machine learning models that automatically learn from terminal behavior patterns and network conditions without requiring manual configuration. The traffic steering controller autonomously generates and updates routing rules based on observed data, enabling the system to self-adapt to changing conditions while reducing the complexity of manual network management.
2Reliability
If proactive congestion control with prediction is implemented, then future network states can be anticipated and prevented, but the system complexity and computational requirements increase
Solution Approach 1:
The system uses machine learning models to predict future network congestion states and terminal behavior patterns before they actually occur. By anticipating potential congestion scenarios in advance, the system can proactively adjust traffic steering rules to prevent congestion rather than reactively responding to it, thereby improving reliability while the predictive algorithms manage the complexity of control decisions.
3Productivity
If terminal-side information is collected and used for predictions, then network optimization can be improved, but the amount of data to be processed and transmitted increases
Solution Approach 1:
The system extracts only the most relevant features from terminal-side information for prediction purposes, rather than processing all available data. The machine learning models are trained to identify and utilize key behavioral patterns and network parameters that most significantly impact traffic optimization, thereby improving network optimization efficiency while minimizing the data volume that needs to be collected, transmitted, and processed.
Data Source
AI summary
Systems and methods for optimizing network traffic are disclosed. In one embodiment, a system for optimizing the performance of a plurality of networks includes a first terminal device, a traffic steering controller, and a terminal traffic steering agent. The traffic steering controller may be configured to receive user profiles and network performance metrics and create traffic steering rules based on the user profiles and network metrics. The terminal traffic steering agent may be configured to receive traffic steering rules from the traffic steering controller and direct a virtual network switch based on the traffic steering rules. The virtual network switch may be configured to receive network traffic at a virtual network interface and, based on directions received from the traffic steering controller, forward the network traffic to a physical interface of at least one of the plurality of networks.