TCP-WBQ: a backlog-queue-based congestion control mechanism for heterogeneous wireless networks

Area of Science:

• Computer Science
• Networking

Background:

• Conventional Transmission Control Protocol (TCP) variants struggle in heterogeneous wireless networks due to random packet loss and high latency.
• Inaccurate congestion degree estimation on high-latency links hinders fine-grained control for existing TCP protocols.

Purpose of the Study:

• To propose a novel sender-side TCP variant, TCP-WBQ (Wireless Backlog Queue), designed for enhanced performance in heterogeneous wireless networks.
• To accurately identify congestion degrees and effectively manage packet loss and latency oscillations.

Main Methods:

• Constructs a backlog-queue model based on congestion window dynamics.
• Deduces model bounds to differentiate between non-congestion, random packet loss, and latency oscillations.
• Implements sender-side congestion degree detection and adaptive congestion window adjustment schemes.

Main Results:

• TCP-WBQ accurately detects congestion degrees, outperforming conventional TCP variants.
• The proposed method effectively shields against random packet loss and latency time oscillations.
• Simulations demonstrate efficient bandwidth utilization and high performance in single and multiple bottleneck scenarios.

Conclusions:

• TCP-WBQ offers a superior tradeoff between high throughput and congestion avoidance in heterogeneous wireless networks.
• The algorithm achieves competitive fairness and high performance, addressing limitations of existing TCP protocols.