Proactive Congestion Control Research Articles

Proactive Congestion Control by Fast Optical Switching Within 1-ms Delay at Mobile Backhaul Utilizing Traffic Data Volume

Proactive Congestion Control by Fast Optical Switching Within 1-ms Delay at Mobile Backhaul Utilizing Traffic Data Volume

COER: A Network Interface Offloading Architecture for RDMA and Congestion Control Protocol Codesign

RDMA (Remote Direct Memory Access) networks require efficient congestion control to maintain their high throughput and low latency characteristics. However, congestion control protocols deployed at the software layer suffer from slow response times due to the communication overhead between host hardware and software. This limitation has hindered their ability to meet the demands of high-speed networks and applications. Harnessing the capabilities of rapidly advancing Network Interface Card (NIC) can drive progress in congestion control. Some simple congestion control protocols have been offloaded to RDMA NIC to enable faster detection and processing of congestion. However, offloading congestion control to the RDMA NIC faces a significant challenge in integrating the RDMA transport protocol with advanced congestion control protocols that involve complex mechanisms. We have observed that reservation-based proactive congestion control protocols share strong similarities with RDMA transport protocols, allowing them to integrate seamlessly and combine the functionalities of the transport layer and network layer. In this paper, we present COER, an RDMA NIC architecture that leverages the functional components of RDMA to perform reservations and completes the scheduling of congestion control during the scheduling process of the RDMA protocol. COER facilitates the streamlined development of offload strategies for congestion control techniques, specifically proactive congestion control, on RDMA NIC. We use COER to design offloading schemes for eleven congestion control protocols, which we implement and evaluate using a network emulator with a cycle-accurate RDMA NIC model that can load MPI programs. The evaluation results demonstrate that the architecture of COER does not compromise the original characteristics of the congestion control protocols. Compared to a layered protocol stack approach, COER enables the performance of RDMA networks to reach new heights.

DemePro: DEcouple packet Marking from Enqueuing for multiple services with PROactive congestion control

Most of current Data Center Network (DCN) protocols leverage Explicit Congestion Notification (ECN) for congestion control to maintain low latency and high throughput. However, the majority of them only consider single-queue scenario in each switch port, making their performance inferior in multiple-service multiple-queue scenario. To this end, we propose DemePro, a DCN scheme for multiple-service multiple-queue scenario. In face of congestion signal, DemePro also leverages ECN for congestion notification, while decouples packet marking from enqueuing, in order to ensure fairness among multiple services. We also question the effectiveness of the congestion signal derived from the single threshold for port buffer queue length, via a set of experiments. Then, DemePro utilizes multiple thresholds for proactive congestion control, and packets are encapsulated to carry congestion extent information, in order to notify TCP senders for precise and fine-grained congestion control. Experiments show that DemePro has better performance than MQ-ECN [1] which is currently the best DCN scheme for multiple-service multiple-queue scenario, and DemePro can also well guarantee the fairness.

A Receiver-Driven Transport Protocol With High Link Utilization Using Anti-ECN Marking in Data Center Networks

Existing reactive or proactive congestion control protocols are hard to simultaneously achieve ultra-low latency and high link utilization across all workloads ranging from delay-sensitive flows to bandwidth-hungry ones in datacenter networks. We present an Anti-ECN (Explicit Congestion Notification) Marking Receiver-driven Transport protocol called AMRT, which achieves both near-zero queueing delay and full link utilization by reasonably increasing sending rate in the case of under-utilization. Specifically, switches mark the ECN bit of data packets once detecting spare bandwidth. When receiving the anti-ECN marked packet, the receiver generates the corresponding marked grant to trigger more data packets. The testbed and simulation experiments show that AMRT effectively reduces the average flow completion time (AFCT) by up to 42% and improves the link utilization by up to 38% over the state-of-the-art receiver-driven transmission schemes.

PayDebt: Reduce Buffer Occupancy Under Bursty Traffic on Large Clusters

The average/tail Flow Completion Times (FCTs) are critical to many datacenter applications. Congestion control plays a central role in optimizing FCT. Inappropriate congestion control can exacerbate buffer occupancy, thus hurting the flow performance. Our observations are that current approaches are too aggressive in injecting packets into underlying networks. Instead of handling buffer explosion afterward, we reduce buffer occupancy in the first place. We propose PayDebt, a novel and readily-deployable proactive congestion control protocol. At its heart, a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">debt</i> mechanism provides bandwidth coordination between the already-buffered and the forthcoming packets. We evaluate PayDebt both in a testbed and large-scale simulations. The buffer occupancy can be decreased by up to 8.0×-35.9× compared to DCQCN and Homa.

ABQ: Active Buffer Queueing in Datacenters

Congestion control is always an active research field for guaranteeing the Quality of Services (QoS) of datacenter networks (DCNs). However, the current end-to-end TCP design enables both senders and receivers not to directly and precisely obtain the congestion information, incurring inaccurate or non-real-time adjustments. Moreover, switches that act as fundamental primitives in DCN are usually used for reactive congestion feedbacks (e.g., Ack and ECN), leading to underused for proactive congestion control. In this article, we propose a novel active queue management (AQM) scheme called Active Buffer Queueing (ABQ) which leverages active buffer queueing of DCN switches to achieve both high-throughput and loss-free transmissions. When the traffic pattern in DCN is changed intensely in flow size, number and interval, the feedback information to TCP end hosts are imprecise. Unlike other AQMs, the key idea of ABQ is to adjust the flow rate (or packet pace) directly in the switch according to the real-time congestion state. We explain the design rationale behind ABQ and present simulation results of its performance. Finally, we discuss the implementation on the NetFPGA.

RecFlow: SDN-based receiver-driven flow scheduling in datacenters

Datacenter applications (e.g., web search, recommendation systems, and social networking) are designed to have a high fanout for the purpose of achieving scalable performance. Frequent fabric congestion (e.g., due to incast, imperfect hashing) is a corollary of such a design. This is true even when the network utilization is low. Such fabric congestion exhibits both temporal as well as spatial (intra-rack and inter-rack) variations. There exist two basic design paradigms which are used to address this issue. Current solutions lie somewhere between the two. On one hand we have arbiter based approaches where senders poll a centralized arbiter and collectively obey global scheduling decisions. On the other end of the spectrum, we have self adjusting end point based approaches where senders independently adjust transmission rate based on network congestion. The former incurs greater overhead, compared to the latter which trades off complexity for sub-optimality. Our work seeks a middle ground - optimality of arbiter based approaches with the simplicity of self adjusting end point based approaches. Our key design principle is that since the receiver has complete information regarding the flows destined for it, rather than having a centralized arbiter schedule flows or the senders making independent scheduling decisions, the receiver can orchestrate the various flows destined for it. Since multiple receivers may be using a bottleneck link, datapath visibility should be used to ensure fair sharing of the bottleneck capacity between receivers with minimum overhead. We propose RecFlow, which is a receiver-based proactive congestion control scheme. RecFlow employs OpenFlow provided path visibility to track changing bottlenecks on the fly. It spaces TCP acknowledgements to prevent traffic bursts and ensure that no receiver exceeds its fair share of the bottleneck capacity. The goal is to reduce buffer overflows while maintaining fairness among flows and high link utilization. Using extensive simulation results and real testbed evaluation, we show that compared to the state-of-the-art, RecFlow achieves up to 6× improvement in the inter-rack scenario and 1.5× in the intra-rack scenario while sharing the link capacity fairly between all flows.

TCP proactive congestion control for east-west traffic: The marking threshold

Various extensions of TCP/IP have been proposed to reduce network latency; examples include Explicit Congestion Notification (ECN), Data Center TCP (DCTCP) and several proposals for Active Queue Management (AQM). Combining these techniques requires adjusting various parameters, and recent studies have found that it is difficult to do so while obtaining both high throughput performance and low latency. This is especially true for mixed use data centres that host both latency-sensitive applications and high-throughput workloads with east–west traffic such as Hadoop.This paper studies the difficulty in configuration, and characterises the problem as related to ACK packets. Such packets cannot be set as ECN Capable Transport (ECT), with the consequence that a disproportionate number of them are dropped. The same issue can affect other non-ECT-capable traffic that may co-exist on the network. We explain how this behavior adversely affects throughput, and propose a small change to the way that non-ECT-capable packets are handled in the network switches. Using NS–2 simulation, we demonstrate robust performance for modified AQMs on a Hadoop cluster, maintaining full throughput while reducing latency by 85%. We also demonstrate that commodity switches with shallow buffers are able to reach the same throughput as deeper buffer switches.Finally, we explain how both TCP using ECN and DCTCP can achieve the best performance using a simple marking threshold, in contrast to the current preference for relying on AQMs to mark packets. Overall, we provide recommendations to network equipment manufacturers, cluster administrators and the whole industry on how best to combine high-throughput and latency-sensitive workloads. This article is an extension of our previous work [1], which was published in Proceedings of the 19th IEEE International Conference on Cluster Computing (CLUSTER 2017).

A Comprehensive Review of Reactive and Proactive Congestion Control Methodologies

A Comprehensive Review of Reactive and Proactive Congestion Control Methodologies

A distributed and quiescent max-min fair algorithm for network congestion control

Given the higher demands in network bandwidth and speed that the Internet will have to meet in the near future, it is crucial to research and design intelligent and proactive congestion control and avoidance mechanisms able to anticipate decisions before the congestion problems appear. Nowadays, congestion control mechanisms in the Internet are based on TCP, a transport protocol that is totally reactive and cannot adapt to network variability because its convergence speed to the optimal values is slow. In this context, we propose to investigate new congestion control mechanisms that (a) explicitly compute the optimal sessions’ sending rates independently of congestion signals (i.e., proactive mechanisms) and (b) take anticipatory decisions (e.g., using forecasting or prediction techniques) in order to avoid the appearance of congestion problems. In this paper we present B-Neck, a distributed optimization algorithm that can be used as the basic building block for the new generation of proactive and anticipatory congestion control protocols. B-Neck computes proactively the optimal sessions’ sending rates independently of congestion signals. B-Neck applies max-min fairness as optimization criterion, since it is very often used in traffic engineering as a way of fairly distributing a network capacity among a set of sessions. B-Neck iterates rapidly until converging to the optimal solution and is also quiescent. The fact that B-Neck is quiescent means that it stops creating traffic when it has converged to the max-min rates, as long as there are no changes in the sessions. B-Neck reacts to variations in the environment, and so, changes in the sessions (e.g., arrivals and departures) reactivate the algorithm, and eventually the new sending rates are found and notified. To the best of our knowledge, B-Neck is the first distributed algorithm that maintains the computed max-min fair rates without the need of continuous traffic injection, which can be advantageous, e.g., in energy efficiency scenarios. This paper proposes as novelty two theoretical contributions jointly with an in-depth experimental evaluation of the B-Neck optimization algorithm. First, it is formally proven that B-Neck is correct, and second, an upper bound for its convergence time is obtained. In addition, extensive simulations were conducted to validate the theoretical results and compare B-Neck with the most representative competitors. These experiments show that B-Neck behaves nicely in the presence of sessions arriving and departing, and its convergence time is in the same range as that of the fastest (non-quiescent) distributed max-min fair algorithms. These properties encourage to utilize B-Neck as the basic building block of proactive and anticipatory congestion control protocols.

Loss Rate Control Mechanism for Fan-in-burst Traffic in Data Center Network

Congestion scenarios in Data Center Network (DCN) arise due to burst traffic and cause packet drop to take place thus reducing the overall throughput. Flow scheduling techniques in DCN do not address well the network congestions. Congestion control techniques uses congestion notifications from network core to deal with congestion scenario. Software defined networking techniques use link load information in access switches to react to congestion scenarios. Both the mechanisms operate on post-congestion scenario to deal with sustained burst traffic. In fat tree topology based DCN architectures proactive measures for handling burst traffic at lower layers will be more beneficial. In this paper, we implement traffic shaping mechanism in the edge switch at source that act proactively and prevent the propagation of ill effects due to sustained burst. Further, we evaluate its impact on the overall packet loss and delay. The entire DCN is simulated using Colored Petri Nets (CPN). The packet loss rates observed at the receiver edge switch for various flow patterns reveals cent percent packet transfer which signifies the effectiveness of the proactive congestion control mechanism.

Application-Based Congestion Control Policy for the Communication Channel in VANETs

This letter proposes a novel proactive congestion control policy for vehicular ad-hoc networks, in which each vehicle's communication parameters are adapted based on their individual application requirements. Contrary to other approaches, where transmission resources tend to be assigned based on system-level performance metrics, the technique proposed in this paper aims to individually satisfy the target application performance of each vehicle, while globally minimising the channel load to prevent channel congestion.

Edge-to-edge proactive congestion control for aggregated traffic

We present Ping Trunking, a novel edge-to-edge management technique that can provide soft service guarantees to aggregate traffic streams without requiring any special support at the core of the network. Our proposal is designed to work over aggregated flows that bundle a varying number of user flows for common treatment between two nodes in a network. To regulate the user data transmission rate, a Vegas-like management connection is established between the two edges of each aggregate. This control connection injects control packets into the network to probe its congestion level. Thanks to this managing, Ping Trunking is able to fairly share the network bandwidth among competing aggregates in accordance with their subscribed target rates. In addition, it does not cause undesired sharp variations in the transmission rates of handled aggregates and avoids packet losses at the core nodes. We demonstrate analytically and through simulation experiments the effectiveness of our technique.

PI-PD Controller for Adaptive and Robust Active Queue Management for Internet Congestion Control

Since random early detection (RED) was proposed in 1993, many active queue management (AQM) algorithms have been proposed to support better end-to-end Transmission Control Protocol (TCP) congestion control. In this article, the authors introduce and analyze a feedback control model of the TCP/AQM dynamics. Then they suggest the concept of an AQM algorithm that can detect and avoid congestion proactively. Finally, they propose the proportional-integral (PI) proportional-derivative (PD) controller using proportional-integral-derivative (PID) feedback control to overcome the reactive control behavior of existing AQM proposals. The PI-PD controller is able to provide proactive congestion avoidance and control using an adaptive congestion indicator and a control function. A comparative simulation study under a variety of network environments shows that the PI-PD controller outperforms RED and the PI controller in terms of the queue length dynamics, the packet loss rates, and the link utilization.

PAQM: an adaptive and proactive queue management for end‐to‐end TCP congestion control

AbstractTwo functions, the congestion indicator (i.e. how to detect congestion) and the congestion control function (i.e. how to avoid and control congestion), are used at a router to support end‐to‐end congestion control in the Internet. Random early detection (RED) (IEEE/ACM Trans. Networking 1993; 1(4):397–413) enhanced the two functions by introducing queue length averaging and probabilistic early packet dropping. In particular, RED uses an exponentially weighted moving average (EWMA) queue length not only to detect incipient congestion but also to smooth the bursty incoming traffic and its resulting transient congestion. Following RED, many active queue management (AQM)‐based extensions have been proposed. However, many AQM proposals have shown severe problems with detection and control of the incipient congestion adaptively to the dynamically changing network situations. In this paper, we introduce and analyse a feedback control model of TCP/AQM dynamics. Then, we propose the Pro‐active Queue Management (PAQM) mechanism, which is able to provide proactive congestion avoidance and control using an adaptive congestion indicator and a control function under a wide range of traffic environments. The PAQM stabilizes the queue length around the desired level while giving smooth and low packet loss rates and high network resource utilization. Copyright © 2004 John Wiley &amp; Sons, Ltd.

Proactive congestion control for high speed networks

The next generation high speed networks which are characterized by large bandwidth propagation delay products are expected to support applications with diverse traffic characteristics and Quality of Service (QoS) requirements. Since statistical multiplexing is employed in these networks, congestion control and avoidance mechanisms are required for the delivery of QoS required by various applications. A number of reactive flow control schemes have been developed in the past for best effort applications. It has been well established that these schemes are not effective for networks with large bandwidth propagation delay products as the feedback information generated in the network can be outdated by the time it reaches the sources. In this paper, this problem (referred to as the bandwidth mismatch problem) is investigated and a proactive feedback (PF) based flow control scheme which attempts to eliminate the bandwidth mismatch problem by generating and transmitting early feedback messages is introduced. With cooperation among the network nodes and sources, the PF based flow control scheme can significantly increase the QoS delivered to the best effort applications. In addition to the increased level of delivered QoS, fair rate allocations to the best effort applications are also possible since the efficiency of the scheme becomes independent of the length of the propagation delays, once the bandwidth mismatch is eliminated.