Flow Sensing-Based Congestion Detection for D2D Streaming on a 5G gNB.

Area of Science:

• Wireless Communications
• Network Engineering
• Data Streaming

Background:

• High-quality streaming in device-to-device (D2D) communications requires monitoring of encoding, decoding, and flow rates.
• Traffic flow imbalance and congestion negatively impact D2D streaming performance.

Purpose of the Study:

• To propose a novel algorithm, FS-CDA, for detecting and controlling flow imbalance in D2D streaming.
• To enhance D2D streaming service performance in 5G wireless mobile networks by preventing high congestion rates.

Main Methods:

• Developed the flow sensing-based congestion detecting algorithm (FS-CDA).
• Implemented transmission rate monitoring, rate adjustment, and underflow/overflow sensing.
• Derived individual weights for streaming flow versions for performance evaluation.

Main Results:

• The proposed FS-CDA algorithm effectively controls traffic flow rates influenced by bandwidth, bit errors, and radio interference.
• Simulation results demonstrate superior performance compared to existing methods in average congestion control ratio, PSNR, and average throughput.

Conclusions:

• FS-CDA significantly enhances D2D streaming service performance in 5G networks.
• The algorithm provides effective congestion control and rate management for optimized streaming.