SLBN: A Scalable Max-min Fair Algorithm for Rate-Based Explicit Congestion Control

Mozó, Alberto; López-Presa, José Luis; Anta, Antonio Fernndez

doi:10.1109/nca.2012.40

Cited by 2 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although MMF can be achieved with low-cost algorithms in single-path networks [17]- [20], that is not the case when dealing with multipath, where algorithms require solving a series of linear programming problems with non-negligible computational cost [28], [33], [42]. Thus, UMMF [27] has been proposed as a relaxation of max-min fairness that can be achieved by purely combinatorial algorithms, which makes this fairness criterion a better candidate for use in multipath networks.…”

Section: Discussionmentioning

confidence: 99%

“…These algorithms require per-connection information maintained at the routers. To overcome this problem, s-PERC [19] and SLBN [20] have been proposed as stateless versions of previous algorithms. One of the most interesting properties of this type of algorithm is that the convergence time to the optimal rates depends on the number of bottleneck levels (in the notation of [18]) and the RTT, but not on the capacity of the links, as in window-based algorithms which progressively increase the transmission rate at the sources.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Upward Max-Min Fairness in Multipath High-Speed Networks

2021

Self Cite

View full text Add to dashboard Cite

To take full advantage of the available network capacity, connections need to be able to use multiple paths to route their packets. Max-min fairness (MMF) can be effectively applied to single-path networks, but computing MMF rates in multipath networks requires solving a series of linear programming (LP) problems with high computational cost. Thus, a relaxation of MMF has been proposed, namely, upward max-min fairness (UMMF), which can be solved by simple combinatorial algorithms. Current proposals carry out incremental approximations emulating the waterfilling algorithm, which inherently establishes a dependency between the time required to achieve the optimal solution and the capacity of the links. Thus, the more capacity the network has, the less efficient the algorithms are. We defined the concept of the saturation level as the basis for the computation of fair shares. We developed the first centralized algorithm based on this concept, which we call c-SLEN. Unlike its predecessors, its convergence time does not depend on network capacity, and it does not incur link oversaturation. Based on c-SLEN, we derived d-SLEN, a distributed protocol that does not need to maintain per-subflow information in routers and guarantees constant processing time for control packets, making it a good candidate for practical use. Finally, through extensive simulations, we showed that d-SLEN is faster, lighter, and more accurate than its counterparts. Owing to its accuracy and convergence speed, it is able to maintain the size of link queues at minimal values at all times, thus proactively avoiding network congestion.INDEX TERMS ERC algorithms, multipath networks, upward max-min fairness, high-speed networks, proactive congestion control.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Upward Max-Min Fairness in Multipath High-Speed Networks

2021

Self Cite

View full text Add to dashboard Cite

show abstract

A Distributed Algorithm to Calculate Max-Min Fair Rates Without Per-Flow State

JoseLavanya¹,

IbanezStephen²,

AlizadehMohammad³

et al. 2019

Proc. ACM Meas. Anal. Comput. Syst.

View full text Add to dashboard Cite

Most congestion control algorithms, like TCP, rely on a reactive control system that detects congestion, then marches carefully towards a desired operating point (e.g. by modifying the window size or adjusting a rate). In an effort to balance stability and convergence speed, they often take hundreds of RTTs to converge; an increasing problem as networks get faster, with less time to react. This paper is about an alternative class of congestion control algorithms based on proactive-scheduling: switches and NICs "pro-actively" exchange control messages to run a \em distributed algorithm to pick "explicit rates for each flow. We call these Proactive Explicit Rate Control (PERC) algorithms. They take as input the routing matrix and link speeds, but not a congestion signal. By exploiting information such as the number of flows at a link, they can converge an order of magnitude faster than reactive algorithms. Our main contributions are (1) s-PERC ("stateless" PERC), a new practical distributed PERC algorithm without per-flow state at the switches, and (2) a proof that s-PERC computes exact max-min fair rates in a known bounded time, the first such algorithm to do so without per-flow state. To analyze s-PERC, we introduce a parallel variant of standard waterfilling, 2-Waterfilling. We prove that s-PERC converges to max-min fair in 6N rounds, where N is the number of iterations 2-Waterfilling takes for the same routing matrix. We describe how to make s-PERC practical and robust to deploy in real networks. We confirm using realistic simulations and an FPGA hardware testbed that s-PERC converges 10-100x faster than reactive algorithms like TCP, DCTCP and RCP in data-center networks and 1.3--6x faster in wide-area networks (WANs). Long flows complete in close to the ideal time, while short-lived flows are prioritized, making it appropriate for data-centers and WANs.

show abstract

SLBN: A Scalable Max-min Fair Algorithm for Rate-Based Explicit Congestion Control

Cited by 2 publications

References 22 publications

Upward Max-Min Fairness in Multipath High-Speed Networks

Upward Max-Min Fairness in Multipath High-Speed Networks

A Distributed Algorithm to Calculate Max-Min Fair Rates Without Per-Flow State

Contact Info

Product

Resources

About