BBR: Congestion-Based Congestion Control

Cardwell, Neal; Cheng, Yuchung; Gunn, C. Stephen; Yeganeh, Soheil Hassas; Jacobson, Van

doi:10.1145/3012426.3022184

Cited by 765 publications

(546 citation statements)

References 12 publications

Supporting

Mentioning

470

Contrasting

Unclassified

Order By: Relevance

“…It is well-known that regardless of how many links a connection traverses, "from TCP's viewpoint an arbitrarily complex path behaves as a single link with the same round-trip time and bottleneck rate" [2,13]. What is less well-understood is the fact that not all bottleneck links are of equal importance.…”

Section: Transmission Of Bottleneck Informationmentioning

confidence: 99%

“…In this initial experiment, we test the hypothesis that TCP congestion-controlled networks behave as predicted by the Theory of Bottleneck Ordering. We carry out three experiments based on Google's B4 network (Figure 8), each with a different number of BPG levels (from 2 to 4 levels), and evaluate whether three well-known congestion control algorithms-BBR [2], Cubic [10] and Reno [7]-are able to recognize the bottleneck structure. We choose these three algorithms as representatives of two broad classes of algorithms.…”

Section: On the Bottleneck Structure Of Tcp Networkmentioning

confidence: 99%

“…For the past three decades since the first congestion control algorithm was implemented as part of the TCP protocol [13], many different variations have been designed, implemented and extensively put into practice to help address these two objectives. In this context, it is well-known that regardless of the complexity of a communication path, the performance of a flow is uniquely determined by the capacity of its bottleneck link and its end-to-end round trip time (RTT) [2,13]. This funnel view has steered much of the research towards the singlebottleneck characterization problem (e.g., [3], [13], [2], [25]), leading to the implicit assumption that bottlenecks in a network have a flat structure, and inadvertently hiding the need for a better understanding of bottlenecks in distributed networks.…”

Section: Introductionmentioning

confidence: 99%

“…We then use three well-known congestion control algorithms to validate whether bottlenecks and flows in real networks behave as predicted by the theory. To get a broader sense of how the theory performs, we choose representatives from two widely used classes of congestion control algorithms: BBR from the class of congestion-based algorithms [2], and Cubic and Reno [7,10] from the class of loss-based additive-increase/multiplicative-decrease (AIMD) algorithms. The key contributions of this paper can be summarized as follows:…”

Section: Introductionmentioning

confidence: 99%

“…insight provides a new formal explanation to understand the reason why such a class of algorithms may perform better [2]. (Sections 3.1 and 3.2.)…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

On the Bottleneck Structure of Congestion-Controlled Networks

Ros-Giralt

Bohara

Yellamraju

et al. 2019

Proc. ACM Meas. Anal. Comput. Syst.

View full text Add to dashboard Cite

In this paper, we introduce the Theory of Bottleneck Ordering, a mathematical framework that reveals the bottleneck structure of data networks. This theoretical framework provides insights into the inherent topological properties of a network in at least three areas: (1) It identifies the regions of influence of each bottleneck; (2) it reveals the order in which bottlenecks (and flows traversing them) converge to their steady state transmission rates in distributed congestion control algorithms; and (3) it provides key insights into the design of optimized traffic engineering policies. We demonstrate the efficacy of the proposed theory in TCP congestion-controlled networks for two broad classes of algorithms: Congestion-based algorithms (TCP BBR) and loss-based additive-increase/multiplicative-decrease algorithms (TCP Cubic and Reno). Among other results, our network experiments show that: (1) Qualitatively, both classes of congestion control algorithms behave as predicted by the bottleneck structure of the network; (2) flows compete for bandwidth only with other flows operating at the same bottleneck level; (3) BBR flows achieve higher performance and fairness than Cubic and Reno flows due to their ability to operate at the right bottleneck level; (4) the bottleneck structure of a network is continuously changing and its levels can be folded due to variations in the flows' round trip times; and (5) against conventional wisdom, low-hitter flows can have a large impact to the overall performance of a network.

show abstract

Section: Transmission Of Bottleneck Informationmentioning

confidence: 99%

Section: On the Bottleneck Structure Of Tcp Networkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…insight provides a new formal explanation to understand the reason why such a class of algorithms may perform better [2]. (Sections 3.1 and 3.2.)…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

On the Bottleneck Structure of Congestion-Controlled Networks

Ros-Giralt

Bohara

Yellamraju

et al. 2019

Proc. ACM Meas. Anal. Comput. Syst.

View full text Add to dashboard Cite

show abstract

References

2020

Network Traffic Engineering

View full text Add to dashboard Cite

SDCCP: Control the network using software‐defined networking and end‐to‐end congestion control

Lin

Liao

Wang

et al. 2020

Concurrency and Computation

View full text Add to dashboard Cite

The Internet of Things is becoming widely popular in the past decade, which comes with huge amount of data. These magnanimous data, stored in data centers, put forward the new demand for the efficient management of the network. In this article, we propose Software-Defined Congestion Control Plane (SDCCP), a hybrid network control architecture that aims to fully utilize the network while avoiding congestion. SDCCP is based on Software-Defined Networking and CCP, in which the controller collects the network statistics and specifies the behavior of the end-to-end hosts by sending feedback or modifying their transport layer parameters directly. It can also be used to mitigate Distributed Denial of Service attacks and other security problems. In addition, we propose FCA, a Feedback-based Congestion Avoidance algorithm running on SDCCP, which adapts the congestion window based on the feedback from the remote controller. We evaluate SDCCP and FCA in Mininet and the result shows that FCA can achieve high network utilization while keeping the queue length of the routers in a low level. Also, FCA is robust to noncongestion loss, and outperforms other algorithms at high loss rate.

show abstract

BBR: Congestion-Based Congestion Control

Cited by 765 publications

References 12 publications

On the Bottleneck Structure of Congestion-Controlled Networks

On the Bottleneck Structure of Congestion-Controlled Networks

References

SDCCP: Control the network using software‐defined networking and end‐to‐end congestion control

Contact Info

Product

Resources

About