No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Performance Evaluation of Explicit Congestion Notification (ECN) in IP Networks
Abstract
This memo presents a performance study of the Explicit Congestion Notification (ECN) mechanism in the TCP/IP protocol using our implementation on the Linux Operating System. ECN is an end-to-end congestion avoidance mechanism proposed by (6) and incorporated into RFC 2481(7). We study the behavior of ECN for both bulk and transactional transfers. Our experiments show that there is improvement in throughput over NON ECN (TCP employing any of Reno, SACK/FACK or NewReno congestion control) in the case of bulk transfers and substantial improvement for transactional transfers. A more complete pdf version of this document is available at: http://www7.nortel.com:8080/CTL/ecnperf.pdf This memo in its current revision is missing a lot of the visual representations and experimental results found in the pdf version.
... Since ECN-Capable routers on the path are able to report the impending congestion before the congestion really happens, ECN is highly expected as a decisive solution to avoid dropping packets. The effectiveness of using ECN has been shown in the previous works [6], [7]. Nevertheless, other works [8]- [11] show that the deployment of ECN on the Internet is much slower than its expectation even though it was standardized as RFC 3168 [4] in 2001. ...
... Performance of ECN and its effectiveness was evaluated by computer simulation [6] and by testbed [7] as it was proposed. However, as hardware and software have greatly changed in network environments in the last two decades, the effectiveness of ECN should be verified again at a testbed close to the current Internet. ...
... However, the throughput efficiency for a particular ECN-Capable TCP (a single iperf) connection is not always improved if it is unfortunately "ecn marked" by the ECN-Capable switch. Effectiveness of using ECN can be found in [6], [7], [27], and the detailed figures and discussion are omitted here due to space limitations. ...
ECN, as a decisive approach for TCP congestion control, has been proposed for many years. However, its deployment on the Internet is much slower than expected. In this paper, we investigate the state of the deployment of ECN (Explicit Congestion Notification) on the Internet from a different viewpoint. We use the data set of web domains published by Alexa as the hosts to be tested. We negotiate an ECN-Capable and a Not ECN-Capable connections with each host and collect all packets belonging to the connections. By analyzing the header fields of the TCP/IP packets, we dig out the deployment rate, connectivity, variation of round-trip time and time to live between the Not ECN-Capable and ECN-Capable connections as well as the rate of IPv6-Capable web servers. Especially, it is clear that the connectivity is different from the domains (regions on the Internet). We hope that the findings acquired from this study would incentivize ISPs and administrators to enable ECN in their network systems.
... The advantages of ECN have been verified in [2], [5] and [12]. In [2], one set of tests employs a single WAN link. ...
... In [5], Linux boxes with experimental ECN implementations were used to verify that ECN improves both bulk and transactional TCP traffic when compared with TCP NewReno and SACK (Selective Acknowledgment, [9]); the gain was more significant for short connections (socalled mice). In [12], fast Cisco routers were employed. ...
... In this way, we were able to check multiple cases and multiple levels of congestion, from no congestion (no packet drops, no marks) to a heavily congested network, with multiple packets dropped or marked by routers. Also, even if congestion was so severe that RED could degenerate to TailDrop [5], we still expected ECN to work no worse that non-ECN. The results are presented in Fig. 4. The three curves present the relative increase of the aggregate number of marks (and drops), aggregate goodput of all sessions and average queue length. ...
Explicit Congestion Notification is a recently proposed congestion avoidance scheme for IP networks. ECN uses marking packets instead of dropping them in case of incipient congestion. Avoiding packet drops saves network bandwidth and allows congestion signals to be propagated faster. In the paper, we present the results of detailed simulations that we have performed to assess the gain TCP applications could achieve from using ECN. We analyze both a single congested link and a more complex configuration. We also discuss head marking, a modification of router's behavior that allows to faster propagate congestion signals.
... The idea looks promising and its advantages have been verified in [4]. However, in the context of IP networks, which currently support mainly best-effort mode with very limited QoS support, some questions may be asked. ...
... The advantages of ECN have been verified in [2], [4]. Nonetheless, they are also checked in the paper. ...
... Two modes of operation were examined (as in [4]): bulk and transactional transfers. In the first case, a long, 2 Mbyte file was downloaded by clients. ...
The paper discusses Explicit Congestion Notification (ECN) in the context of the commercial Internet. It presents possible threats to ECN and shows the gain malicious ECN sources could achieve from ECN subverting. Besides, the paper analyses the relationships between ECN and Service Level Agreements (SLAs).
... Many authors have pointed out that marking provides more information about the congestion state than packet dropping [69,32], and ECN has been proven to be a better way to deliver the congestion signal and exhibits a better performance [26,69,16]. ...
... Many authors have pointed out that marking provides more information about the congestion state than packet dropping [69,32], and ECN has been proven to be a better way to deliver the congestion signal and exhibits a better performance [26,69,16]. ...
... While some research has confirmed the benefits of ECN (Floyd, 1994;Pentikousis et al., 2001;Pentikousis and Badr, 2001;Salim and Ahmed, 2000), the benefits reported usually relate to improvements in goodput or throughput and are a consequence of packets travelling only once over a bottle neck link. Other research has suggested that ECN inherits problems associated with the underlying AQM which is generally RED (Floyd and Jacobson, 1993). ...
... In total 15 combinations of ECN and non-ECN flows were tested. Between 1 and 8 streams were used for each test which is comparable to the methodology used in RFC2884 where congestion levels of low, medium and high corresponded to 1, 5 and 10 flows respectively (Salim and Ahmed, 2000). ...
This paper outlines the results of an experimental investigation into the dynamics of RFC3168 based Explicit Congestion Notification (ECN) using congestion avoidance methods that are available in commercial routing equipment. Careful analysis of the packet exchanges and queuing conditions provide insight into TCP ECN's behaviour on production equipment using largely default parameters. We find that the relative performance of ECN and non-ECN enabled TCP depends heavily on the prevailing network conditions. The use of ECN leads to improved goodput and in most instances throughput. However, this improvement comes at the expense of volatile queues leading to instability and degraded performance under certain circumstances.
... This factor is important because the congestion duration can be shortened, the packet loss reduced and the network resources better used. Traditional TCP's congestion control and avoidance algorithms [14] are powerful but not enough to provide good service in a lot of network conditions since they handle the network as if it were a black box [9]. The goal of our work is to create a new congestion control model, which we call Open Box Protocol (OBP), using router collaboration to identify the network resources along the path and to provide this information to end systems. ...
... They are briefly explained as follows. The ECN [9], [10] and the Quick-Start [11] are two solutions that use the router collaboration to address the congestion control problem. With the ECN, the router collaboration is done by detecting congestion situations and by informing the end systems about this situation. ...
In this paper we propose a new explicit congestion control approach, Open Box Protocol (OBP). The OBP gives sources the capacity
to look inside the network and to make their congestion control decisions, based, not only on packets loss or packets delay,
but also based on another kind of information. For example, the most restricted interface capacity, the available bandwidth,
the RTT variations or the presence of heterogeneous transmission means. In the paper we describe the OBP and discuss its evaluation
results based on ns-2 simulations. The results show the OBP´s capacity to provide traffic sources with the required information,
in static or dynamic network scenarios, and to make correct and quick congestion control decisions. Also, it is visible that
the OBP avoids the full queue problem and tries to keep the queues near zero occupation.
... DOI: 10.1587/transinf.2020NTP0002 [6] and [7] evaluated the effectiveness of ECN and the benefits of using ECN are summarized in [8]. [9] used a new response strategy in ECN and got improved throughput and reduced fluctuations. ...
In this paper, we used the data set of domain names Global Top 1M provided by Alexa to analyze the effectiveness of Fallback in ECN. For the same test server, we first negotiate a connection with Not-ECN-Capable, and then negotiate a connection with ECN-Capable, if the sender does not receive the response to ECN-Capable negotiation from the receiver by the end of retransmission timeout, it will enter the Fallback state, and switch to negotiating a connection with Not-ECN-Capable. By extracting the header fields of the TCP/IP packets, we confirmed that in most regions, connectivity will be slightly improved after Fallback is enabled and Fallback has a positive effect on the total time of the whole access process. Meanwhile, we provided the updated information about the characteristics related to ECN with Fallback in different regions by considering the geographical region distribution of all targeted servers.
... • With ECN, the AQM can signal immediately, and the sender can smooth the signals-it knows its own RTT, which it can use as the appropriate smoothing time [4] or it can choose to respond without smoothing when appropriate, e.g. at flow start. 3) ECN also offers the obvious latency benefit of near-zero congestion loss, which primarily benefits short flows, as shown by Salim & Ahmed [55]. This removes retransmission and time-out delays and the head-of-line blocking that a loss can cause when a transport with ordered delivery (like TCP) carries a multiplex of streams. ...
On the Internet, sub-millisecond queueing delay and capacity-seeking have traditionally been considered mutually exclusive. We introduce a service that offers both: Low Latency Low Loss Scalable throughput (L4S). When tested under a wide range of conditions emulated on a testbed using real residential broadband equipment, queue delay remained both low (median 100--300 $\mu$s) and consistent (99th percentile below 2 ms even under highly dynamic workloads), without compromising other metrics (zero congestion loss and close to full utilization). L4S exploits the properties of `Scalable' congestion controls (e.g., DCTCP, TCP Prague). Flows using such congestion control are however very aggressive, which causes a deployment challenge as L4S has to coexist with so-called `Classic' flows (e.g., Reno, CUBIC). This paper introduces an architectural solution: `Dual Queue Coupled Active Queue Management', which enables balance between Scalable and Classic flows. It counterbalances the more aggressive response of Scalable flows with more aggressive marking, without having to inspect flow identifiers. The Dual Queue structure has been implemented as a Linux queuing discipline. It acts like a semi-permeable membrane, isolating the latency of Scalable and `Classic' traffic, but coupling their capacity into a single bandwidth pool. This paper justifies the design and implementation choices, and visualizes a representative selection of hundreds of thousands of experiment runs to test our claims.
... This is referred to as Tail Drop. There are a number of drawbacks related to the Tail Drop mechanism which gave rise to smarter congestion control mechanisms in routers such as Random Early Detection (RED) [2]. ...
In current TCP/IP networks, TCP mainly relies on packet drops as an indication of congestion. With the emergence of Explicit Congestion Notification (ECN), TCP can detect the congestion through packets with marked bits instead of dropped packets. ECN thus reduces the unnecessary delay due to packet drops, especially for bandwidth-delay sensitive connections. However , ECN still relies on the receiver for congestion indication, which incurs in a round trip time (RTT) before the sender can react to the congestion. In network environments with the high bandwidth-delay product (BDP), this could be problematic. In this paper, we present an enhanced ECN mechanism for the early detection of congestion using P4 programming. In our proposed scheme, the sender does not have to wait for the receiver to indicate congestion because switches in the network can indicate congestion. We evaluate our scheme comparing our proposed solution with standard TCP and TCP with the conventional ECN. Experimental results show that our scheme outperforms ECN more as the RTT grows.
... AQM calculates the current value of aql according to the number of packets then compares it with the set threshold. When the aql value arrives to defined threshold, all the incoming packets that arrive at the router buffer are dropped with main factor called the probability in order to preventing router buffer overflow and becomes full [24][25][26]. Enormous methods for congestion control, such as AGRED [22], EAGRED [27], DGRED [14], and FLGRED have been built based on AQM. ...
This paper proposes the Stabilized (DGRED) method for congestion detection at the router buffer. This method aims to stabilize the average queue length between allocated minthre_shold and doublemaxthre_shold positions to increase the network performance. The SDGRED method is simulated and compared with Gentle Random Early Detection (GRED) and Dynamic GRED active queue management methods. This comparison is built on different important measures, such as dropping probability, throughput, average delay, packet loss, and mean queue length for packets. The evaluation aims to identify which method presents better simulation performance measurement results when non-congestion or congestion situations occur at the router buffers in congestion control. The results show that at high packet arrival probability, the proposed algorithm helps provide lesser queue length values, delayed time, and packet loss compared with current methods. Furthermore, SDGRED generates adequate throughput at high packet arrival probability.
... Despite early evidence of its positive impact [27], a succession of unfortunate incidents stalled ECN deployment for 15 years: some firewalls treated all TCP/ECN connection attempts as port scanning attacks [14]; TCP/ECN connection attempts were mistakenly discarded by certain home router models and one popular model crashed [30]; and when routinely wiping the Diffserv field between networks, a bug wiped the IP/ECN field too. ...
After ECN was first added to IP in 2001, it was hit by a succession of deployment problems. Studies in recent years have concluded that path traversal of ECN has become close to universal. In this article, we test whether the performance enhancement called ECN++ will face a similar deployment struggle as did base ECN. For this, we assess the feasibility of ECN++ deployment over mobile as well as fixed networks. In the process, we discover bad news for the base ECN protocol: contrary to accepted beliefs, more than half the mobile carriers we tested wipe the ECN field at the first upstream hop. All packets still get through, and congestion control still functions, just without the benefits of ECN. This throws into question whether previous studies used representative vantage points. This article also reports the good news that, wherever ECN gets through, we found no deployment problems for the "++" enhancement to ECN. The article includes the results of other in-depth tests that check whether servers that claim to support ECN actually respond correctly to explicit congestion feedback. Those interested can access the raw measurement data online.
... Active queue management methods have been proposed to improve the network performance (Baklizi et al., 2013, Abdel-jaber et al., 2008. Many researchers have been proposed an AQM method, such as (Kiruthiga and Raj, 2014, Das et al., 2013, Singh andBalveer, 2013), which were proposed to overcome the limitations of the DT method discussed earlier (Bitorika et al., 2004, Salim andAhmed, 2000). Enormous methods for congestion control have been built as AQM, such as Random Early Detection (RED) (Floyd and Jacobson, 1993), Adaptive Random Early Detection (ARED) (Floyd et al., 2001), Random Exponential Marking (REM) (Athuraliya et al., 2001, Lapsley andLow, 1999) , BLUE (Feng et al., 1999, Feng et al., 2002, Stochastic Fair BLUE (SFB) (Feng et al., 2001), Gentle Random Early Detection (GRED) (Floyd, 2000), Dynamic Random Early Drop (DRED) (Aweya et al., 2001), Stabilized Random Early Drop (SRED) (Ott et al., 1999), Fuzzy BLUE (Yaghmaee and AminToosi, 2003), Fuzzy Exponential Marking (FEM) (Chrysostomou et al., 2003), Decbit (Ramakrishnan and Raj, 1988), Enhanced Random Early Detection(ENRED) (Alshimaa et al., 2014), Adaptive Neuro Fuzzy Inference System (ANFIS) (Kusumawardani, 2013), AGRED , and DGRED (Baklizi et al., 2013) . ...
Congestion control methods are continuously linked with the rapid advances in Internet and network technology. Congestion generally occurs when the amount of packets arriving at the router buffer cannot be accommodated. This paper proposed an Enhanced Adaptive Gentle Random Early Detection (Enhanced AGRED) method based on Adaptive Gentle Random Early Detection (AGRED) method in order to detect the congestion in early stage before the router buffer overflows by enhancing the parameter setting of Queue Weight (Qw). The Enhanced AGRED is simulated and compared with the AGRED and Gentle Random Early Detection (GRED) methods. The simulation results for the proposed Enhanced AGRED, GRED and AGRED methods are carried out by varying the variable of packet arrival probability to create different congestion/non-congestion scenarios. During the congestion, the simulation results reveals that Enhanced AGRED offers marginally better performance results than GRED and AGRED, with regard to mean queue length, average queuing delay and packet loss probability due to overflow. Therefore, the results prove that Enhanced AGRED is an effective method in controlling congestion router buffers of networks. Whereby, improve networks performance.
... In parameter-based early congestion control methods, the congestion is monitored and controlled at an early stage before the router overflows, using a set of parameters. These methods are formally referred to as Active Queue Management (AQM) methods, which were proposed to overcome the limitations of the DT method discussed earlier [9,10]. Enormous methods for congestion control have been built as AQM, such as Random Early Detection (RED) [11], Adaptive Random Early Detection (ARED) [12], Random Exponential Marking (REM) [13,14] , BLUE [15,16], Stochastic Fair BLUE (SFB) [17], Gentle Random Early Detiction (GRED) [18], Dynamic Random Early Drop (DRED) [19], Stabilised Random Early Drop (SRED) [20], DRED [21,22] Fuzzy BLUE [23], Fuzzy Exponential Marking (FEM) [24], Decbit [25] and Adaptive Gentle Random Early Detection. ...
... La figure ci-dessous schématise le principe de fonctionnement de cette politique. Cette politique de gestion de files d'attente est notamment utilisée pour les flots de type TCP en guise de mécanisme de contrôle de congestion : en effet, la méthode ECN (Early Congestion Notification) [57,88] utilise le même principe adopté par RED pour indiquer à une source TCP d'adapter son débit d'émission en fonction de la charge du réseau. L'aspect négatif de RED est qu'il s'agit d'un mécanisme qui ne permet pas de gérer la priorité des paquets, ce qui pourrait largement affecter des flots à haute priorité demandant une importante qualité de service. ...
The notion of quality of service (QoS) is very essential each time we want to transfer the information with a maximum of reliability. The main metrics of QoS are the throughput, the delay, the jitter and the loss rate. In addition, the various natures of applications leads to heterogeneity of the data : elastic flows require a guaranteed throughput while non-elastic flows require a guaranteed delay. The scheduling algorithms that use a proportional differentiation make possible to fairly share the resources between applications, according to their needs. However, the proportionality was carried out according to only one quality metric (throughput, delay or loss rate). Thus, we designed and propose a scheduling algorithm using proportional differentiation, founded on the power function which is the ratio of two criteria: the delay and the throughput. PSP scheduler (Power as a Scheduling Parameter) offers a fairly resource sharing between the applications that have different needs. In this way, no class suffers from the famine phenomenon. The implementation and the tests of the scheduler were carried out by the ns-2 simulation tool.
... AQM first calculates the value of aql then compares it with the given threshold. When the aql value is greater than the threshold value, all packets arriving at the buffer are dropped with the probability of preventing router buffer overflow [16][17][18]. ...
The present paper proposes the Dynamic Gentle Random Early Detection (DGRED) algorithm for early stage congestion detection at the router buffer. Generally, the proposed DGRED algorithm depends on the stability of the average queue length at a specific level between allocated minimum and maximum threshold values, with the aim to improve the network performance. The DGRED algorithm is simulated and compared with the most known Active Queue Management Early Detection (RED) algorithm and two of its variants, namely, Gentle RED and Adaptive GRED. This comparison was conducted based on different performance measures, such as mean queue length, throughput, average queuing delay, packet loss, and dropping probability for packets. The comparison aimed to identify the algorithm that offers better performance measurement results under either non-congestion or congestion situation at the router buffers. The acquired results show that the proposed algorithm contributes in providing lesser queue length, delayed queuing, and packet loss probability compared with the existing algorithms when high packet arrival probability appears, that is, (> 0.63). Furthermore, DGRED generates adequate throughput when the packet arrival probability value is high.
... Congestion is indicated by a single packet drop in non-ECN-Capable TCP. ECN mechanisms have been tested and shown to improve throughput over NON-ECN Reno TCP for bulk transfer as well as transactional transfer [59], due to the need for fewer retransmissions. ...
Stability of the Internet today depends largely on cooperation between end hosts that employ TCP (Transmission Control Protocol) protocol in the transport layer, and network routers along an end-to-end path. However, in the past several years, various types of traffic, including streaming media applications, are increasingly deployed over the Internet. Such types of traffic are mostly based on UDP (User Datagram Protocol) and usually do not employ neither end-to-end congestion norflow control mechanism, or else very limited. Such applications could unfairly consume greater amount of bandwidth than competing responsive flows such as TCP traffic. In this manner, unfairness problem and congestion collapse could occur. To avoid substantial memory requirement and complexity, fair Active Queue Management (AQM) utilizing no or partial flow state information were proposed in the past several years to solve these problems. These schemes however exhibit several problems under different circumstances.This dissertation presents two fair AQM mechanisms, BLACK and AFC, that overcome the problems and the limitations of the existing schemes. Both BLACK and AFC need to store only a small amount of state information to maintain and exercise its fairness mechanism. Extensive simulation studies show that both schemes outperform the other schemes in terms of throughput fairness under a large number of scenarios. Not only able to handle multiple unresponsive traffic, but the fairness among TCP connections with different round trip delays is also improved. AFC, with a little overhead than BLACK, provides additional advantages with an ability to achieve good fairness under a scenario with traffic of diff21erent sizes and bursty traffic, and provide smoother transfer rates for the unresponsive flows that are usually transmitting real-time traffic.This research also includes the comparative study of the existing techniques to estimate the number of active flows which is a crucial component for some fair AQM schemes including BLACK and AFC. Further contribution presented in this dissertation is the first comprehensive evaluation of fair AQM schemes under the presence of various type of TCP friendly traffic.
... Floyd mentioned this problem as an open issue of ECN in[10]. Salim et al.[11]experimentally evaluated ECN performance in networks with a small number of flows. Their results showed that, for two competing flows, the relative advantage for an ECN-capable flow increases by up to sixty percent as the congestion level increases. ...
In this paper, we study the performance of Active Queue Management (AQM) algorithms in Explicit Congestion Notification (ECN) capable networks. We show that Random Early Detection (RED) with ECN can cause severe fluctuations in queue length by dropping ECN-capable packets, and that bandwidth allocation is unfair, when ECN-capable and non-ECN-capable traffic coexist. To resolve these problems, we propose the Hybrid Fair Queueing (HFQ) algorithm. We also provide guidelines for setting the HFQ parameters, based on a control-theoretic analysis. We show that the HFQ algorithm outperforms other AQM algorithms using ns-2 simulations.
... The advantages of ECN is avoidance (or reduction) of packet loss due to quick reaction to congestion. The reduction of packet drops in turn reduces latency and jitter, and thus, the number of re-transmitted packets [29]. However, ECN requires Active Queue Management 5 and per-packet processing. ...
... In EETCP network itself participate in controlling its own resource utilization by making use of ECN algorithm [10] idea is to mark rather than drop packets at the router. EETCP further make use of improved RED [6] random early detection which drops packet randomly as a function of the average queue size. ...
Increasing popularity of real time traffic application has lead to increase in non TCP traffic over the internet. These non TCP applications does not provide adequate congestion control compared to TCP application sharing same network thus available bandwidth is shared unfairly and may lead to stall of TCP traffic. M any TCP-Friendly protocols have been proposed that can share the available bandwidth with TCP traffic fairly and in such a way that Non TCP flow have same throughput as TCP one when are under same network conditions. These schemes have different characteristic and are best suited for different network conditions. In this paper we discuss various available TCP friendly schemes and discuss their features.
... Cooperating end systems should then use this IP mark as a signal that the network is congested and slow down the throughput. This is known as Explicit Congestion Notification (ECN) [53]. ...
The current congestion control mechanism used in TCP has difficulty reaching full utilization on high speed links, particularly on wide-area connections. For example, the packet drop rate needed to fill a Gigabit pipe using the present TCP protocol is below the currently achievable fiber optic error rates. HighSpeed TCP was recently proposed as a modification of TCP's congestion control mechanism to allow it to achieve reasonable performance in high speed wide-area links. In this research, simulation results showing the performance of HighSpeed TCP and the impact of its use on the present implementation of TCP are presented. Network conditions including different degrees of congestion, different levels of loss rate, different degrees of bursty traffic and two distinct router queue management policies were simulated. The performance and fairness of HighSpeed TCP were compared to the existing TCP and solutions for bulk-data transfer using parallel streams.
... The TCP sender will then reduce its congestion window. Thus, the use of ECN is believed to provide performance benefits [32], [43]. RFC 3168 [9] also places requirements on intermediate routers (e.g., active queue management and setting of the CE bit(s) in the IP header to indicate congestion). ...
... Explicit Congestion Notification (ECN) [3] was suggested as an effective method to drop packets. Experimental study in [4] has shown the performance advantages of ECN for TCP short transfers. ...
We present a new end-to-end transport protocol called Multi-path Transmission Control Protocol (M/TCP) and its robust acknowledgement (ACK) schemes. M/TCP is designed as an alternative TCP option to improve reliability and performance of today's Internet. M/TCP allows a sender to simultaneously transmit data via multiple controlled paths to the same destination. The protocol requires no modification in IP layer. In order to establish multiple paths, however, two endpoints communicating through M/TCP need to be subscribed to multiple ISPs 1 . Congestion control and error recovery in M/TCP are developed based on those in TCP. The ACK schemes introduced in this paper provide a mechanism to improve M/TCP performance over Internets with high packet-loss in an ACK channel. We compare performance of M/TCP with TCP Reno implementation using ns2 simulator. Simulation results show that M/TCP can achieve significantly higher throughput than TCP Reno in the presence of error model on forward and reverse paths.
... They are briefly explained as follows. The ECN [8], [9] and the Quick-Start [10] are two solutions that use the router collaboration to address the congestion control problem. With the ECN, the router collaboration is done by detecting congestion situations and by informing the end systems about this situation. ...
In this paper we propose Per-flow Bandwidth Distribution in Routers (BDR). This is an approach which enables an efficient use of the available bandwidth in high or low speed networks. BDR gives routers capacities to detect the number of active flows that are crossing the routers' output interface. This is done without needing to maintain any per-flow state information. In the paper we describe BDR and discuss its evaluation results based on ns-2 simulations. The results show that BDR has capacities to discover the number of active flows and has capacities to distribute the resources between all flows.
... Active queue management methods have been proposed to improve the network performance (Baklizi et al., 2013, Abdel-jaber et al., 2008. Many researchers have been proposed an AQM method, such as (Kiruthiga and Raj, 2014, Das et al., 2013, Singh andBalveer, 2013), which were proposed to overcome the limitations of the DT method discussed earlier (Bitorika et al., 2004, Salim andAhmed, 2000). Enormous methods for congestion control have been built as AQM, such as Random Early Detection (RED) (Floyd and Jacobson, 1993), Adaptive Random Early Detection (ARED) (Floyd et al., 2001), Random Exponential Marking (REM) (Athuraliya et al., 2001, Lapsley andLow, 1999) , BLUE (Feng et al., 1999, Feng et al., 2002, Stochastic Fair BLUE (SFB) (Feng et al., 2001), Gentle Random Early Detection (GRED) (Floyd, 2000), Dynamic Random Early Drop (DRED) (Aweya et al., 2001), Stabilized Random Early Drop (SRED) (Ott et al., 1999), Fuzzy BLUE (Yaghmaee and AminToosi, 2003), Fuzzy Exponential Marking (FEM) (Chrysostomou et al., 2003), Decbit (Ramakrishnan and Raj, 1988), Enhanced Random Early Detection(ENRED) (Alshimaa et al., 2014), Adaptive Neuro Fuzzy Inference System (ANFIS) (Kusumawardani, 2013), AGRED , and DGRED (Baklizi et al., 2013) . ...
The developments in computer networks in recent days such as the internet have increased rapidly. Connections of these networks necessitate resources in order to send their data to their prospective destinations. Further, the connections require high speed router buffers which they route data in high speed. Congestion is one of the main issues that occur at the router buffer cause deterioration of the network performance, i.e. increasing average waiting time, decreasing throughput, etc. Gentle Random Early Detection (GRED) is one of the known congestion control algorithms proposed to detect congestion before the router buffer overflows. In fact, GRED improves the setting of the parameters for the maximum threshold position (max threshold) at the router buffer and the maximum value for the packet dropping probability (D max ). This paper proposes an Adaptive GRED algorithm that detects congestion at router buffers in an preliminary stage, and enhances the parameters setting of the max threshold and the D max . During congestion, the simulation results reveal that the Adaptive GRED drops fewer packets than GRED, and it marginally offers better performance results than that of GRED.
Emerging wireless systems target to provide multi-Gbps data rates for each user, which can be achieved by utilizing ultra-wide channels available at mmWave, terahertz, and lightwave frequencies. In contrast to the traditional spectrum below 6 GHz, these high-frequency bands raise many issues, complicating their usage. For example, because of high signal attenuation and blockage by obstacles, the data rates in a high-frequency band may quickly vary by several orders of magnitude. This peculiarity is often considered a challenge for modern transport layer protocols, such as Transmission Control Protocol (TCP) or Quick UDP Internet Connections (QUIC). Their key component is the Congestion Control Algorithm (CCA), which tries to determine a data sending rate that maximizes throughput and avoids network congestion. Many recent studies show that the performance of the existing CCAs significantly degrades if mobile devices communicate with high-frequency bands and propose some solutions to address this problem. The goal of this survey is twofold. First, we classify the reasons for poor TCP & QUIC performance in high-frequency bands. Second, we comprehensively review the solutions already designed to solve these problems. In contrast to existing studies and reviews that mainly focus on the comparison of various CCAs, we consider solutions working at different layers of the protocol stack, i.e., from the transport layer down to the physical layer, as well as cross-layer solutions. Based on the analysis, we conclude the survey with recommendations on which solutions provide the highest gains in high-frequency bands.
Today, usage of internet is growing exponentially. Congestion detection and avoidance algorithms are the major issues in TCP/IP. Earlier, packet drops are only source of congestion indication, but it leads to loss of throughput. Active Queue Management (AQM) can detect congestion before the queue overflows and informs the end hosts to respond congestion. It allows gateways to drop packets when average queue is greater than maximum threshold and marks the packets otherwise. Explicit Congestion Notification (ECN) mechanism marks the packets when the average queue size is between predefined thresholds and improves the throughput of a network. But it depends on the end hosts to respond to congestion. So there is a possibility of misbehavior by sender to increase its congestion window, even if the receiver correctly signals about congestion. So misbehaving ECN sender flow obtains more throughput than the normal ECN-enabled flows. We present an Extended ECN mechanism that enables a router to mark packets and the receiver to signal congestion to the sender without trusting the sender whether it has responded congestion or not. Our improved mechanism is robust in detection and prevention of this misbehaving sender in network and compatible with ECN and TCP/IP mechanisms.
To detect network congestion, TCP typically relies on detecting packet loss. While this is an effective approach for maintaining high throughput for bulk data transfers, a better approach for interactive, time-sensitive, or loss-sensitive traffic would be to detect congestion prior to packet loss. Explicit Congestion Notification (ECN) is a congestion avoidance strategy that makes use of Active Queue Management to allow TCP endpoints to detect congestion without a corresponding packet drop. This congestion detection strategy is particularly useful for delay-sensitive traffic where packet retransmissions can lead to noticeable delays for the user. In this paper, we present an implementation of ECN as an addition to the TCP protocols in ns-3. We have modified all TCP variants currently in ns-3 to work with this new addition. To validate our work we tested all TCP variants and compared our implementation's behavior to previous work.
Explicit Congestion Notification (ECN) is a TCP/IP extension that can avoid packet loss and thus improve network performance. Though standardized in 2001, it is barely used in today's Internet. This study, following on previous active measurement studies over the past decade, shows marked and continued increase in the deployment of ECN-capable servers, and usability of ECN on the majority of paths to such servers. We additionally present new measurements of ECN on IPv6, passive observation of actual ECN usage from flow data, and observations on other congestion-relevant TCP options (SACK, Timestamps and Window Scaling). We further present initial work on burst loss metrics for loss-based congestion control following from our findings.
Reducing the delay of web traffic has been an important issue in Internet works. However, it is difficult to reduce delay while improving throughput. This is because, in order to reduce delay queue size should be less but to improve throughput queue size should be more. In our modification, web traffic and ECN marked packets are dropped only when high congestion occurs at router. We propose a solution to improve response time as well as the number of packets transmitted of web traffic without affecting throughput of bottleneck link. We applied this solution to RED algorithm and tested the performance and efficiency of modified RED algorithm as compared to the original RED algorithm.
Larger round trip time of a packet and its acknowledgement may expire the retransmission timeout timer. Packets not yet reaching their destinations are then thought as lost and retransmitted. As a result, the traffic with retransmitted packets increases and then round trip time also increases. The procedure is repeated. The Internet may become congested. In this paper, we propose a method based on the fuzzy theory to adapt retransmission timeout timer. Simulation results show our proposed scheme outperforms over the original TCP retransmission scheme. Keyword: Congestion Collapse, Fuzzy Theory
In current telecommunication, how to communicate on the network is one of the important issues. Handover will occur for free movement of mobile node. Cellular network has covered a wide range but has a relatively narrow bandwidth. WLAN has a wider bandwidth but a relatively narrow range of services available. Heterogeneous handover will occur for free movement of mobile node. During the handover process will have inevitably transport delay and network congestion. Currently research has progressed with Freeze-TCP and TCP Westwood. But wireless congestion and handover congestion cannot distinguish. In this paper, we propose that receiver (mobile node) determines the occurrence of the handover. Receiver received ECN marked packet and readjusts the window size according to the handover. Than sender receive handover option field and ECN marked ACK, then it sends dummy packet. In this way, problems can be solved by changing TCP header and ECN. The result of this paper is use Qualnet
We propose a model of router queue buffering with FAST TCP flows in dumbbell networks. Using this model, we analysis the transient dynamic of router queue buffering with different scenarios. It is found that, network settings as well as FAST TCP parameters can influence router queue buffering transient performance. We provide simulation results to verify the theoretical results.
TCPpsilas congestion control provides remarkable stability to the media streaming. In the course of implementation of the TCP protocol, in order to achieve optimal throughput TCP led to the delay and reduce the quality of the media streaming transmission. In this paper, we introduced dynamic tuning of the send-side buffer ways to minimize this latency.The experiment proved that the method introduced by this article can effectively reduce the latency caused by TCP layer, hence reduce the end-to-end latency of TCP connection.
As an enhancement mechanism for the end-to-end congestion control, active queue management (AQM) can keep smaller queuing delay and higher throughput by purposefully dropping the packets at the intermediate routers. Comparing with RED algorithm, although the PI (proportional-integral) or PID (proportional-integral-differential) controller for AQM improves the stability, It is very difficulty in selecting a group of the parameters of the controller to guarantee the transient performance. Moreover, the case is even worse especially in rapidly changing and complex conditions, such as stochastically accessed users, unpredicted nonresponsive traffic and hard limitations on the network resources. In this paper, we propose a self-tuning controller for AQM router supporting TCP with ECN based on neural compensator. The effectiveness of integrated performance of the controller is demonstrated by the simulation results
Congestion control is widely used in the Internet to prevent congestion collapse. Because data are inherently bursty, routers are provisioned with large buffers to absorb this burstiness and to achieve high link utilization. At the same time, large buffers lead to high queuing delays at congested routers. RED (Random Early Detection) was introduced to relieve this problem so as to achieve high link utilization and low queue delay. Several adaptive techniques have also been proposed to allow for better parameter adjustment under different situation settings. However, parameter adjustment approaches in these techniques are usually based on an assumption that there exists a known combination of optimal parameter setting based on which techniques are to adjust to. Whereas, optimality of the setting very much depends on circumstantial factors which cannot be universally true, and thus the ensuing adjustment may not be even beneficial. In this paper, we propose an Autonomous Random Early Detection (AURED) technique to allow for a complete autonomous adjustment process without having to assume the aforementioned association. By tuning the packet drop probability variable according to the performance variation between two consecutive sampling periods, this technique does not require a target setting value to adapt to, thus allowing for more flexibility to accommodate for various situations.
This document proposes a new mechanism for TCP and Stream Control
Transmission Protocol (SCTP) that can be used to recover lost
segments when a connection’s congestion window is small. The
"Early Retransmit" mechanism allows the transport to reduce, in certain
special circumstances, the number of duplicate acknowledgments
required to trigger a fast retransmission. This allows the transport
to use fast retransmit to recover segment losses that would otherwise
require a lengthy retransmission timeout.
Delivering congestion signals is essential to the performance of networks. Current TCP/IP networks use packet losses to signal congestion. Packet loss not only reduces TCP performance, but also adds large delay. Explicit congestion notification (ECN) delivers a faster indication of congestion and has better performance. However, current ECN implementations mark the packet from the tail of the queue. In this paper, we propose the mark-front strategy to send an even faster congestion signal. We show that mark-front strategy reduces buffer size requirement, improves link efficiency and provides better fairness among users. Simulation results that verify our analysis are also presented.
A window-based flow control is a sort of feedback-based congestion control mechanisms, and has been widely used in current
TCP/IP networks. In the feedback-based congestion control, feedback delay of the congestion information from the bottleneck
node is one of the important elements that affect the performance. Recently, the use of an explicit congestion notification
(ECN) mechanism as congestion indication from the network to source hosts has been actively discussed in the IETF. In this
paper, we propose an enhanced marking strategy for ECN that can reduce the transfer delay of congestion information of a router
in multiple-hop network environments. At a transit router, the proposed method relays the congestion experienced (CE)-bit
contained in an incoming IP packet to the head-of-line packet with a corresponding flow which is waiting for transmission
in the output buffer. Simulation results showed that ECN with the proposed mark-relay strategy produced a superior performance
in terms of throughput and fairness between TCP flows.
The TCP was designed to work in an environment that is almost link error free. So it supposes that any error during receiving
data is due to congestion. This assumption is not always true, especially in a wireless environment, which has a high BER
when compared to other medium like fiber and twisted pair. This work describes a simple and backward compatible way to increase
the performance of TCP in wireless links. This solution is evaluated in a simulated environment using the NS2 simulator and
is based in informing the TCP whenever the link is under high BER, so it can trigger the proposed algorithm that will hold
the transmission and save the congestion window status for late recovery from the disconnection.
ResearchGate has not been able to resolve any references for this publication.