BBRv2+: Towards balancing aggressiveness and fairness with delay-based bandwidth probing

Fu-rong, Yang; Wu, Qinghua; Li, Zhenyu; Liu, Yanmei; Pau, Giovanni; Xie, Gaogang

doi:10.1016/j.comnet.2022.108789

Cited by 12 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this subsection, we discuss the fairness and friendliness for Antelope. We evaluate the fairness between flows by Jain's fairness index [37], [38], which is calculated as: ((…”

Section: F Fairness and Friendlinessmentioning

confidence: 99%

A Machine Learning-Based Framework for Dynamic Selection of Congestion Control Algorithms

Zhou

Qiu

et al. 2023

IEEE/ACM Trans. Networking

Self Cite

View full text Add to dashboard Cite

Most congestion control algorithms (CCAs) are designed for specific network environments. As such, there is no known algorithm that achieves uniformly good performance in all scenarios for all flows. Rather than devising a one-size-fitsall algorithm (which is a likely impossible task), we propose a system to dynamically switch between the most suitable CCAs for specific flows in specific environments. This raises a number of challenges, which we address through the design and implementation of Antelope, a system that can dynamically reconfigure the stack to use the most suitable CCA for individual flows. We build a machine learning model to learn which algorithm works best for individual conditions and implement kernellevel support for dynamically switching between CCAs. The framework also takes application requirements of performance into consideration to fine-tune the selection based on applicationlayer needs. Moreover, to reduce the overhead introduced by machine learning on individual front-end servers, we (optionally) implement the CCA selection process in the cloud, which allows the share of models and the selection among front-end servers. We have implemented Antelope in Linux, and evaluated it in both emulated and production networks. The results demonstrate the effectiveness of Antelope via dynamic adjusting the CCAs for individual flows. Specifically, Antelope achieves an average 16% improvement in throughput compared with BBR, and an average 19% improvement in throughput and 10% reduction in delay compared with CUBIC.

show abstract

“…In this subsection, we discuss the fairness and friendliness for Antelope. We evaluate the fairness between flows by Jain's fairness index [37], [38], which is calculated as: ((…”

Section: F Fairness and Friendlinessmentioning

confidence: 99%

A Machine Learning-Based Framework for Dynamic Selection of Congestion Control Algorithms

Zhou

Qiu

et al. 2023

IEEE/ACM Trans. Networking

Self Cite

View full text Add to dashboard Cite

show abstract

“…Swift algorithm, which uses delay as a congestion signal to achieve higher fairness and lower retransmission rate. BBRv2 can solve some fairness issues and limitations of the original BBR, but there are still some RTT fairness and intra-protocol fairness issues [22,[30][31][32][33]. Kfoury et al [31] pointed out that BBRv2 s slow ability to identify available bandwidth in network contexts with erratic bandwidth leads to low link utilization.…”

Section: Related Workmentioning

confidence: 99%

Optimization of BBR Congestion Control Algorithm Based on Pacing Gain Model

Yang

Tang

Pan

et al. 2023

Sensors

View full text Add to dashboard Cite

In 2016, Google proposed a congestion control algorithm based on bottleneck bandwidth and round-trip propagation time (BBR). The BBR congestion control algorithm measures the network bottleneck bandwidth and minimum delay in real-time to calculate the bandwidth delay product (BDP) and then adjusts the transmission rate to maximize throughput and minimize latency. However, relevant research reveals that BBR still has issues such as RTT unfairness, high packet loss rate, and deep buffer performance degradation. This article focuses on its most prominent RTT fairness issue as a starting point for optimization research. Using fluid models to describe the data transmission process in BBR congestion control, a fairness optimization strategy based on pacing gain is proposed. Triangular functions, inverse proportional functions, and gamma correction functions are analyzed and selected to construct the pacing gain model, forming three different adjustment functions for adaptive adjustment of the transmission rate. Simulation and real experiments show that the three optimization algorithms significantly improve the fairness and network transmission performance of the original BBR algorithm. In particular, the optimization algorithm that employs the gamma correction function as the gain model exhibits the best stability.

show abstract

iPerfOPS: A Tool for Machine Learning-Based Optimization Through Protocol Selection

Anvari

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

BBRv2+: Towards balancing aggressiveness and fairness with delay-based bandwidth probing

Cited by 12 publications

References 22 publications

A Machine Learning-Based Framework for Dynamic Selection of Congestion Control Algorithms

A Machine Learning-Based Framework for Dynamic Selection of Congestion Control Algorithms

Optimization of BBR Congestion Control Algorithm Based on Pacing Gain Model

iPerfOPS: A Tool for Machine Learning-Based Optimization Through Protocol Selection

Contact Info

Product

Resources

About