Adaptive Wi-Fi Contention Window via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for throughput and low latency in wireless devices leads to contention for the available wireless spectrum, with existing Wi-Fi mechanisms like Binary-Exponential-Backoff (BEB) scaling poorly as the number of interfering transmitters increases, and current algorithms fail to improve performance without compromising fairness in dense deployments.
Innovation Solution
A machine-learning method dynamically adjusts the contention window in Wi-Fi networks by monitoring network traffic, creating a calibration queue, and using a predictive model to set optimal wait times based on system load, ensuring fair access across all devices without requiring changes to client devices or protocols.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If Binary-Exponential-Backoff (BEB) mechanism is used to manage contention, then fairness is maintained, but throughput deteriorates as the number of interfering transmitters increases
Solution Approach 1:
The patent applies dynamics by making the contention window size adaptive rather than static. The system continuously monitors channel conditions and dynamically adjusts the contention window size based on observed traffic patterns and collision rates, allowing the network to optimize performance as the number of transmitters changes
Solution Approach 2:
The patent changes the parameter of contention window size based on system conditions. By monitoring metrics such as collision frequency and channel utilization, the system adjusts the contention window parameter to balance between maintaining fairness and maximizing throughput in dense transmitter environments
2Reliability
If contention window is increased to reduce collisions, then fairness is improved, but latency increases
Solution Approach 1:
The system dynamically adjusts the contention window size based on real-time channel conditions rather than using a fixed large window. This allows the network to maintain fairness when needed while minimizing unnecessary delays when the channel is relatively clear
Solution Approach 2:
The patent implements feedback mechanisms that monitor collision rates and channel utilization, using this information to adjust the contention window size. This feedback loop ensures that fairness is maintained through appropriate window sizing while avoiding excessive latency by reducing the window when conditions permit
3Productivity
If QoS levels are differentiated to prioritize high priority streams, then throughput for priority streams is improved, but fairness to other streams deteriorates
Solution Approach 1:
The patent applies local quality by implementing differentiated contention window sizes for different traffic streams based on their priority levels. High priority streams receive smaller contention windows for faster access, while other streams use larger windows, creating locally optimized access patterns that balance overall fairness with priority handling
Data Source
AI summary
A novel method that dynamically changes the contention window of access points based on system load to improve performance in a dense Wi-Fi deployment is disclosed. A key feature is that no MAC protocol changes, nor client side modifications are needed to deploy the solution. Setting an optimal contention window can lead to throughput and latency improvements up to 155%, and 50%, respectively. Furthermore, an online learning method that efficiently finds the optimal contention window with minimal training data, and yields an average improvement in throughput of 53-55% during congested periods for a real traffic-volume workload replay in a Wi-Fi test-bed is demonstrated.