In data centers, the nature of the composite bursty traffic along with the small bandwidth-delay product and switch buffers lead to several congestion problems that are not handled well by traditional congestion control mechanisms such as TCP. Existing work try to address the problem by modifying TCP to suit the operational nature of data centers. This is practically feasible in private settings, however, in public environments, such modifications are prohibited. Therefore, in this work, we design simple switch-based queue management to deal with such congestion issues adequately. This approach entails no modification to the TCP sender and receiver algorithms which enables easy and seamless deployment in public data centers. We present a theoretical analysis to show the stability and effectiveness of the scheme. We also present, three different real implementations (as a Linux kernel module and as an added feature to OpenvSwitch) and give numerical results from both NS-2 simulation and experiments of real deployment in a small test-bed cluster to show its effectiveness in achieving high throughput overall, a good fairness and short flow completion times for delay-sensitive flows.
Low delay is an explicit requirement for applications such as cloud gaming and video conferencing. Delay-based congestion control can achieve the same throughput but significantly smaller delay than loss-based one and is thus ideal for these applications. However, when a delay- and a loss-based flow compete for a bottleneck, the loss-based one can monopolize all the bandwidth and starve the delay-based one. Fair queuing at the bottleneck link solves this problem by assigning an equal share of the available bandwidth to each flow. However, so far no end host based algorithm to detect fair queui…
Delay-based congestion control algorithms provide higher throughput and stability than traditional loss-based AIMD algorithms, but they are inherently unfair against older connections when the queuing and the propagation delay cannot be measured accurately and independently. This paper presents a novel measurement algorithm whereby fairness between old and new connections is preserved. The algorithm does not modify the dynamics of congestion control, and runs entirely in the server host using locally available information.
Communication networks are time-varying and hence, fair sharing of network resources among the users in such a dynamic environment is a challenging task. In this context, a time-varying network model is designed and the shortest user's route is found. In the designed network model, an end to end window-based congestion control scheme is developed with the help of internal nodes or router and the end user can get implicit feedback (RTT and throughput). This scheme is considered as fair if the allocation of resources among users minimizes overall congestion or backlog in the networks. Window upd…
BBR is a new congestion-based congestion control algorithm proposed by Google. A BBR flow sequentially measures the bottleneck bandwidth and round-trip delay of the network pipe, and uses the measured results to govern its sending behavior, maximizing the delivery bandwidth while minimizing the delay. However, our deployment in geo-distributed cloud servers reveals a severe RTT fairness problem: a BBR flow with longer RTT dominates a competing flow with shorter RTT. Somewhat surprisingly, our deployment of BBR on the Internet and an in-house cluster unearthed a consistent bandwidth disparity a…
Cooperative inter-vehicular applications rely on the exchange of broadcast single-hop status messages among vehicles, called beacons. The aggregated load on the wireless channel due to periodic beacons can prevent the transmission of other types of messages, what is called channel congestion due to beaconing activity. In this paper we approach the problem of controlling the beaconing rate on each vehicle by modeling it as a Network Utility Maximization (NUM) problem. This allows us to formally apply the notion of fairness of a beaconing rate allocation in vehicular networks and to control the …
Cooperative vehicle safety (CVS) systems operate based on broadcast of vehicle position and safety information to neighboring cars. The communication medium of CVS is a vehicular ad-hoc network. One of the main challenges in large scale deployment of CVS systems is the issue of scalability. To address the scalability problem, several congestion control methods have been proposed and are currently under field study. These algorithms adapt transmission rate and power based on network measures such as channel busy ratio. We examine two such algorithms and study their dynamic behavior in time and …
Congestion control is an indispensable component of transport protocols to prevent congestion collapse. As such, it distributes the available bandwidth among all competing flows, ideally in a fair manner. However, there exists a constantly evolving set of congestion control algorithms, each addressing different performance needs and providing the potential for custom parametrizations. In particular, content providers such as CDNs are known to tune TCP stacks for performance gains. In this paper, we thus empirically investigate if current Internet traffic generated by content providers still ad…
Although delay-based congestion control protocols such as FAST promise to deliver better performance than traditional TCP Reno, they have not yet been widely incorporated to the Internet. Several factors have contributed to their lack of deployment. Probably, the main contributing factor is that they are not able to compete fairly against loss-based congestion control protocols. In fact, the transmission rate in equilibrium of delay-based approaches is always less than their fair share when they share the network with traditional TCP-Reno derivatives, that employ packet losses as their congest…
The traditional TCP congestion control mechanism encounters a number of new problems and suffers a poor performance when the IEEE 802.11 MAC protocol is used in multihop ad hoc networks. Many of the problems result from medium contention at the MAC layer. In this paper, I first illustrate that severe medium contention and congestion are intimately coupled, and TCP s congestion control algorithm becomes too coarse in its granularity, causing throughput instability and excessively long delay. Further, we illustrate TCP s severe unfairness problem due to the medium contention and the tradeoff bet…
It is by now well-known that wireless networks with file arrivals and departures are stable if one uses alpha-fair congestion control and back-pressure based scheduling and routing. In this paper, we examine whether ?alpha-fair congestion control is necessary for flow-level stability. We show that stability can be ensured even with very simple congestion control mechanisms, such as a fixed window size scheme which limits the maximum number of packets that are allowed into the ingress queue of a flow. A key ingredient of our result is the use of the difference between the logarithms of queue le…
Recent model-based congestion control algorithms such as BBR use repeated measurements at the endpoint to build a model of the network connection and use it to achieve optimal throughput with low queuing delay. Conversely, applying this model-based approach to Active Queue Management (AQM) has so far received less attention. We propose the new AQM scheduler cocoa based on fair queuing, which adapts the buffer size depending on the needs of each flow without requiring active participation from the endpoint. We implement this scheduler for the Linux kernel and show that it interacts well with th…
In the Tor network, anonymity is achieved through a multi-layered architecture, which comes at the cost of a complex network. Scheduling data in this network is a challenging task and the current approach shows to be incapable of avoiding network congestion and allocating fair data rates. We propose PredicTor, a distributed model predictive control approach, to tackle these challenges. PredicTor is designed to schedule incoming and outgoing data rates on individual nodes of the Tor architecture, leading to a scalable approach. We successfully avoid congestion through exchanging information of …
Congestion control is essential for the stability of the Internet and the corresponding algorithms are commonly evaluated for interoperability based on flow-rate fairness. In contrast, video conferencing software such as Zoom uses custom congestion control algorithms whose fairness behavior is mostly unknown. Aggravatingly, video conferencing has recently seen a drastic increase in use - partly caused by the COVID-19 pandemic - and could hence negatively affect how available Internet resources are shared. In this paper, we thus investigate the flow-rate fairness of video conferencing congestio…
TCP and its variants have suffered from surprisingly poor performance for decades. We argue the TCP family has little hope to achieve consistent high performance due to a fundamental architectural deficiency: hardwiring packet-level events to control responses without understanding the real performance result of its actions. We propose Performance-oriented Congestion Control (PCC), a new congestion control architecture in which each sender continuously observes the connection between its actions and empirically experienced performance, enabling it to consistently adopt actions that result in h…
Production data centers operate under various workload sizes ranging from latency-sensitive mice flows to long-lived elephant flows. However, the predominant load balancing scheme in data center networks, equal-cost multi-path (ECMP), is agnostic to path conditions and performs poorly in asymmetric topologies, resulting in low throughput and high latencies. In this paper, we propose CAFT, a distributed congestion-aware fault-tolerant load balancing protocol for 3-tier data center networks. It first collects, in real time, the complete congestion information of two subsets from the set of all p…
In order to curtail the escalating packet loss rates caused by an exponential increase in network traffic, active queue management techniques such as Random Early Detection (RED) have come into picture. Flow Random Early Drop (FRED) keeps state based on instantaneous queue occupancy of a given flow. FRED protects fragile flows by deterministically accepting flows from low bandwidth connections and fixes several shortcomings of RED by computing queue length during both arrival and departure of the packet. Stochastic Fair Queuing (SFQ) ensures fair access to network resources and prevents a bust…
In today's data center, a diverse mix of throughput-sensitive long flows and delay-sensitive short flows are commonly presented in shallow-buffered switches. Long flows could potentially block the transmission of delay-sensitive short flows, leading to degraded performance. Congestion can also be caused by the synchronization of multiple TCP connections for short flows, as typically seen in the partition/aggregate traffic pattern. While multiple end-to-end transport-layer solutions have been proposed, none of them have tackled the real challenge: reliable transmission in the network. In this p…
We consider two elementary (max-flow and uniform-flow) and two realistic (max-min fairness and proportional fairness) congestion control schemes, and analyse how the algorithms and network structure affect throughput, the fairness of flow allocation, and the location of bottleneck edges. The more realistic proportional fairness and max-min fairness algorithms have similar throughput, but path flow allocations are more unequal in scale-free than in random regular networks. Scale-free networks have lower throughput than their random regular counterparts in the uniform-flow algorithm, which is fa…
Widespread use of computer networks and the use of varied technology for the interconnection of computers has made congestion a significant problem. In this report, we summarize our research on congestion avoidance. We compare the concept of congestion avoidance with that of congestion control. Briefly, congestion control is a recovery mechanism, while congestion avoidance is a prevention mechanism. A congestion control scheme helps the network to recover from the congestion state while a congestion avoidance scheme allows a network to operate in the region of low delay and high throughput wit…
This paper studies the problem of congestion control and scheduling in ad hoc wireless networks that have to support a mixture of best-effort and real-time traffic. Optimization and stochastic network theory have been successful in designing architectures for fair resource allocation to meet long-term throughput demands. However, to the best of our knowledge, strict packet delay deadlines were not considered in this framework previously. In this paper, we propose a model for incorporating the quality of service (QoS) requirements of packets with deadlines in the optimization framework. The sol…
Congestion control algorithms are crucial in achieving high utilization while preventing overloading the network. Over the years, many different congestion control algorithms have been developed, each trying to improve in specific situations. However, their interactions and co-existence has, to date, not been thoroughly evaluated, which is the focus of this paper. Through head-to-head comparisons of representatives from loss-based, delay-based and hybrid types of congestion control algorithms, we reveal that fairness in resources claimed is often not attained, especially when flows sharing a l…
Applications running in geographically distributed setting are becoming prevalent. Large-scale online services often share or replicate their data into multiple data centers (DCs) in different geographic regions. Driven by the data communication need of these applications, inter-datacenter network (IDN) is getting increasingly important. However, we find congestion control for inter-datacenter networks quite challenging. Firstly, the inter-datacenter communication involves both data center networks (DCNs) and wide-area networks (WANs) connecting multiple data centers. Such a network environmen…
The proliferation of cloud data center applications and network function virtualization (NFV) boosts dynamic and QoS dependent traffic into the data centers network. Currently, lots of network routing protocols are requirement agnostic, while other QoS-aware protocols are computationally complex and inefficient for small flows. In this paper, a computationally efficient congestion avoidance scheme, called CECT, for software-defined cloud data centers is proposed. The proposed algorithm, CECT, not only minimizes network congestion but also reallocates the resources based on the flow requirement…
In the context of radio distributed networks, we present a generalized approach for the Medium Access Control (MAC) with fixed congestion window. Our protocol is quite simple to analyze and can be used in a lot of different situations. We give mathematical evidence showing that our performance is tight, in the sense that no protocol with fixed congestion window can do better. We also place ourselves in the WiFi/WiMAX framework, and show experimental results enlightening collision reduction of 14% to 21% compared to the best known other methods. We show channel capacity improvement, and fairnes…
Based on the principle of onion routing, the Tor network achieves anonymity for its users by relaying user data over a series of intermediate relays. This approach makes congestion control in the network a challenging task. As of today, this results in higher latencies due to considerable backlog as well as unfair data rate allocation. In this paper, we present a concept study of PredicTor, a novel approach to congestion control that tackles clogged overlay networks. Unlike traditional approaches, it is built upon the idea of distributed model predictive control, a recent advancement from the …
In this paper, we analyze some local stability and local bifurcation properties of the Proportionally fair, TCP fair, and the Delay-based dual algorithms in the presence of two distinct time delays. In particular, our focus is on the interplay between different notions of fairness, stability, and bifurcation theoretic properties. Different notions of fairness give rise to different non-linear models for the class of Dual algorithms. One can devise conditions for local stability, for each of these models, but such conditions do not offer clear design recommendations on which fairness criteria i…
In this paper, we propose a quality-centric congestion control for multimedia streaming over IP networks, which we refer to as media-TCP. Unlike existing congestion control schemes that adapt a user's sending rate merely to the network condition, our solution adapts the sending rate to both the network condition and the application characteristics by explicitly considering the distortion impacts, delay deadlines, and interdependencies of different video packet classes. Hence, our media-aware solution is able to provide differential services for transmitting various packet classes and thereby, …
The OSU scheme is a rate-based congestion avoidance scheme for ATM networks using explicit rate indication. This work was one of the first attempts to define explicit rate switch mechanisms and the Resource Management (RM) cell format in Asynchronous Transfer Mode (ATM) networks. The key features of the scheme include explicit rate feedback, congestion avoidance, fair operation while maintaining high utilization, use of input rate as a congestion metric, O(1) complexity. This paper presents an overview of the scheme, presents those features of the scheme that have now become common features of…
An explicit rate indication scheme for congestion avoidance in computer and telecommunication networks is proposed. The sources monitor their load and provide the information periodically to the switches. The switches, in turn, compute the load level and ask the sources to adjust their rates up or down. The scheme achieves high link utilization, fair allocation of rates among contending sources and provides quick convergence. A backward congestion notification option is also provided. The conditions under which this option is useful are indicated.
To develop low latency congestion control algorithm for real time taffic has been gained attention recently. RTP Media Congestion Avoidance Techniques (RMCAT) working group was initiated for standard defination. There are three algorithms under this group, Network Assisted Dynamic Adaptation (NADA) proposed by Cisco, Google Congestion Control (GCC) proposed by Google and Self-Clocked Rate Adaptation for Multimedia(SCReAM) proposed by Ericsson. This paper compares and analyses the performance of these algorithms on simulation environment. Results show GCC has well fairness property and performs…
To meet the timing requirements of interactive applications, the no-frills congestion-agnostic transport protocols like UDP are increasingly deployed side-by-side in the same network with congestion-responsive TCP. In cloud platforms, even though the computation and storage is totally virtualized, they lack a true virtualization mechanism for the network (i.e., the underlying data centers networks). The impact of such lack of isolation services, may result into frequent outages (for some applications) when such diverse traffics contend for the small buffers in the commodity switches used in da…
A few months ago, the BitTorrent developers announced that the transfer of torrent data in the official client was about to switch to uTP, an application-layer congestion-control protocol using UDP at the transport-layer. This announcement immediately raised an unmotivated buzz about a new, imminent congestion collapse of the whole Internet. Though this reaction was not built on solid technical foundation, nevertheless a legitimate question remains: i.e., whether this novel algorithm is a necessary building block for future Internet applications, or whether it may result in an umpteenth additi…
Data centers traffic is composed by numerous latency-sensitive "mice" flows, which is consisted of only several packets, and a few throughput-sensitive "elephant" flows, which occupy more than 80% of overall load. Generally, the short-lived "mice" flows induce transient congestion and the long-lived "elephant" flows cause persistent congestion. The network congestion is a major performance inhibitor. Conventionally, the hop-by-hop and end-to-end flow control mechanisms are employed to relief transient and persistent congestion, respectively. However, in face of the mixture of elephants and mic…
Fairness is the significant factor to sustain best effort delivery of network services. Now-a-days, real-time multimedia applications have evolved largely over the Internet. Most of multimedia services are unresponsive during network congestion. Unresponsive traffic streams steal most network bandwidth and starve out some other flows which are responsive during network congestion. In the presence of these unresponsive traffic flows, protection of responsive flows has become a major issue. Many Active Queue Management (AQM) based solutions have been recommended to protect responsive traffic flo…
The increasing number of different, incompatible congestion control algorithms has led to an increased deployment of fair queuing. Fair queuing isolates each network flow and can thus guarantee fairness for each flow even if the flows' congestion controls are not inherently fair. So far, each queue in the fair queuing system either has a fixed, static maximum size or is managed by an Active Queue Management (AQM) algorithm like CoDel. In this paper we design an AQM mechanism (Learning Fair Qdisc (LFQ)) that dynamically learns the optimal buffer size for each flow according to a specified rewar…
We present and investigate a novel and timely application domain for deep reinforcement learning (RL): Internet congestion control. Congestion control is the core networking task of modulating traffic sources' data-transmission rates to efficiently utilize network capacity, and is the subject of extensive attention in light of the advent of Internet services such as live video, virtual reality, Internet-of-Things, and more. We show that casting congestion control as RL enables training deep network policies that capture intricate patterns in data traffic and network conditions, and leverage th…
Congestion control algorithms rely on a variety of congestion signals (packet loss, Explicit Congestion Notification, delay, etc.) to achieve fast convergence, high utilization, and fairness among flows. A key limitation of these congestion signals is that they are either late in feedback or they incur significant overheads. An ideal congestion control must discover any available bandwidth in the network, detect congestion as soon as link utilization approaches full capacity, and react timely to avoid queuing and packet drops, without significant overheads. To this end, this work proposes Scou…
With the increasing demands on interactive video applications, how to adapt video bit rate to avoid network congestion has become critical, since congestion results in self-inflicted delay and packet loss which deteriorate the quality of real-time video service. The existing congestion control is hard to simultaneously achieve low latency, high throughput, good adaptability and fair bandwidth allocation, mainly because of the hardwired control strategy and egocentric convergence objective. To address these issues, we propose an end-to-end statistical learning based congestion control, named Ir…
The recent endeavors of the research community to unite efforts on the design and evaluation of congestion control algorithms have created a growing collection of congestion control schemes called Pantheon. However, the virtual network emulator that comes with the collection has very limited capabilities: it can run flows of only one scheme at once, and the flows cannot have individual network settings, as the topology is point-to-point. This thesis addresses those limitations and presents CoCo-Beholder, a human-friendly emulator providing the popular dumbbell topology of any size, each link o…
Congestion is said to occur in the network when the resource demands exceed the capacity and packets are lost due to too much queuing in the network. During congestion, the network throughput may drop to zero and the path delay may become very high. A congestion control scheme helps the network to recover from the congestion state. A congestion avoidance scheme allows a network to operate in the region of low delay and high throughput. Such schemes prevent a network from entering the congested state. Congestion avoidance is a prevention mechanism while congestion control is a recovery mechanis…
In Future Internet it is possible to change elements of congestion control in order to eliminate jitter and batch loss caused by the current control mechanisms based on packet loss events. We investigate the fundamental problem of adjusting sending rates to achieve optimal utilization of highly variable bandwidth of a network path using accurate packet rate information. This is done by continuously controlling the sending rate with a function of the measured packet rate at the receiver. We propose the relative loss of packet rate between the sender and the receiver (Relative Rate Reduction, RR…
We consider the problem of designing a packet-level congestion control and scheduling policy for datacenter networks. Current datacenter networks primarily inherit the principles that went into the design of Internet, where congestion control and scheduling are distributed. While distributed architecture provides robustness, it suffers in terms of performance. Unlike Internet, data center is fundamentally a "controlled" environment. This raises the possibility of designing a centralized architecture to achieve better performance. Recent solutions such as Fastpass and Flowtune have provided the…
When the stations in an IEEE 802.11 infrastructure Basic Service Set (BSS) employ Transmission Control Protocol (TCP) in the transport layer, this exacerbates per-flow unfair access which is a direct result of uplink/downlink bandwidth asymmetry in the BSS. We propose a novel and simple analytical model to approximately calculate the per-flow TCP congestion window limit that provides fair and efficient TCP access in a heterogeneous wired-wireless scenario. The proposed analysis is unique in that it considers the effects of varying number of uplink and downlink TCP flows, differing Round Trip T…
In this paper we introduced Modified Sized-based Queue Management as a dropping scheme that aims to fairly prioritize and allocate more service to VoIP traffic over bulk data like FTP as the former one usually has small packet size with less impact to the network congestion. In the same time, we want to guarantee that this prioritization is fair enough for both traffic types. On the other hand we study the total link delay over the congestive link with the attempt to alleviate this congestion as much as possible at the by function of early congestion notification. Our M-SQM scheme has been eva…
In heterogeneous networks, achieving congestion avoidance is difficult because the congestion feedback from one subnetwork may have no meaning to source on other other subnetworks. We propose using changes in round-trip delay as an implicit feedback. Using a black-box model of the network, we derive an expression for the optimal window as a function of the gradient of the delay-window curve. The problems of selfish optimum and social optimum are also addressed. It is shown that without a careful design, it is possible to get into a race condition during heavy congestion, where each user wants …
