Thèses sur le sujet « TCP congestion control for hybrid wireless »

The best MPhil thesis in the Faculties of Dentistry, Engineering, Medicine and Science (University of Hong Kong), Li Ka Shing Prize,2010-11.
published_or_final_version
Electrical and Electronic Engineering
Master
Master of Philosophy

Transmission Control Protocol (TCP) designed to deliver seamless and reliable end-to-end data transfer across unreliable networks works impeccably well in wired environment. In fact, TCP carries the around 90% of Internet traffic, so performance of Internet is largely based on the performance of TCP. However, end-to-end throughput in TCP degrades notably when operated in wireless networks. In wireless networks, due to high bit error rate and changing level of congestion, retransmission timeouts for packets lost in transmission is unavoidable. TCP misinterprets these random packet losses, due to the unpredictable nature of wireless environment, and the subsequent packet reordering as congestion and invokes congestion control by triggering fast retransmission and fast recovery, leading to underutilization of the network resources and affecting TCP performance critically. This thesis reviews existing approaches, details two proposed systems for better handling in networks with random loss and delay. Evaluation of the proposed systems is conducted using OPNET simulator by comparing against standard TCP variants and with varying number of hops.

Expanding the global Internet to include mobile devices is an exciting area of current research. Because of the vast size of the Internet, and because the protocols in it are already widely deployed, mobile devices must inter-operate with those protocols. Although most of the incompatiblities with mobiles have been solved, the protocols that deliver data reliably, and that account for the majority of Internet traffic, perform very poorly. A change in location causes a disruption in traffic, and disruption is dealt with by algorithms tailored only for stationary hosts. The Transmission Control Protocol (TCP) is the predominant transport-layer protocol in the Internet. In this thesis, we look at the performance of TCP in mobile environments. We provide a complete explanation for poor performance; we conduct a large number of experiments, simulations, and analyses that prove and quantify poor performance;and we propose simple and scalable solutions that address the limitations.

Les réseaux de télécommunication transparents constituent une étape de développement des réseaux entièrement électroniques. Les technologies de réseau de données actuelles utilisent déjà activement les fibres optiques et les réseaux transparents dans les réseaux centraux, métropolitains et résidentiels. Toutefois, ces réseaux reposent toujours sur la commutation électronique de paquets (EPS) pour le routage des paquets, qui rend obligatoire pour chaque paquet d'avoir une conversion de signal optique à électronique à optique (OEO). D'autre part, la commutation optique de paquets (OPS), qui semblait remplacer le système EPS, promet depuis longtemps des améliorations en termes de performances et de consommation d'énergie en s'éloignant des conversions OEO; cependant, l'absence de buffers optiques pratiques rendait OPS extrêmement vulnérable aux contentions, entraînant une réduction des performances et empêchant de tirer profit des gains de l'OPS. L'objectif de cette recherche est d'étudier la performance des réseaux avec des commutateurs tout optiques et hybrides, tandis que les activités de transmission côté serveur sont régies par des protocoles de contrôle de transport basés sur des algorithmes de contrôle de congestion (TCP CCA). Nous considérons que l'opération OPS pourrait être activée en utilisant un commutateur hybride, c.a.d. une solution au niveau de l'appareil, ainsi que des TCP CCA spécialement conçus, c.a.d. une solution au niveau du réseau, donnant naissance à des réseaux hybrides à commutation de paquets optique (HOPS). Nous étudions les réseaux de centres de données (DCN) de type OPS, HOPS et EPS associés à différentes TCP CCAs en suivant les trois axes de la performance: débit, consommation d'énergie et latence. En ce qui concerne les TCP CCA, nous considérons non seulement les solutions existantes, mais également celles développées. Si Stop-And-Wait (SAW), Selective Acknowledgment (SACK), SACK modifié (mSACK) et Data Center TCP (DCTCP) sont déjà connus, Stop-And-Wait- Longer (SAWL) est présenté ici et conçu pour tirer le meilleur du HOPS DCN. Il est démontré que les solutions de commutateurs hybrides surpassent de manière significative les commutateurs tout optiques sans buffer et atteignent le niveau de commutateurs tout électroniques en termes de débit du réseau. En termes de consommation d'énergie, les solutions hybrides peuvent économiser jusqu'à 4 fois plus d'énergie de la commutation par rapport aux solutions tout électroniques. De plus, les DCN HOPS peuvent atteindre des latences moyennes à l'échelle des microsecondes, dépassant ainsi les EPS et se situant au même niveau que les OPS. La question de l'introduction de classes de service dans HOPS DCN est examinée: on constate que les règles de commutation spécifiques en commutation hybride peuvent améliorer la performance de certaines classes sans pertes significatives d'autres
Transparent optical telecommunication networks constitute a development step from all-electronic networks. Current data network technologies already actively employ optical fibers and transparent networks in the core, metro, and residential area networks. However, these networks still rely on Electronic Packet Switching (EPS) for packets routing, constituting obligatory for each packet optical-to-electronic-to-optical (OEO) signal conversion. On the other hand, Optical Packet Switching (OPS), seemed to be as replacement of EPS, has long promised performance and energy consumption improvements by going away from OEO conversions; however, the absence of practical optical buffers made OPS highly vulnerable to contention, incurring performance reduction, and getting in the way of profiting from OPS gains. The subject of this research lies in the investigation of the performance of OPS networks under all-optical and hybrid switches, while server-side transmission activities are regulated by Transport Control Protocols based on Congestion Control Algorithms (TCP CCAs). We consider that OPS could be enabled by use hybrid switch, i.e. device-level solution, as well by use of specially designed TCP CCAs, i.e. networklevel solution, giving birth to Hybrid Optical Packet Switching (HOPS) networks. We extensively study OPS, HOPS and EPS types of Data Center Networks (DCN) coupled with different TCP CCAs use by following the next three axes of DCN performance: Throughput, Energy Consumption, and Latency. As for TCP CCAs we consider not only existing but also newly developed solutions. If Stop-And-Wait (SAW), Selective Acknowledgment (SACK), modified SACK (mSACK) and Data Center TCP (DCTCP) are already known to the world, StopAnd-Wait-Longer (SAWL) is newly presented and is designed to bring the best out of the HOPS DCN. As a result, it is shown that hybrid switch solutions significantly outperform bufferless all-optical switches and reach the level of all-electronic switches in DCNs in terms of throughput. In terms of energy consumption, hybrid solutions can save up to 4 times on energy on switching compared to all-electronic solutions. As well HOPS DCNs can exhibit microseconds-scale average latencies, surpassing EPS and performing on the level with OPS. The question of the introduction of Classes of Service to HOPS DCN is also investigated: it was found that class-specific switching rules to hybrid switch can ameliorate the performance of certain classes without almost performance loss in others

In this paper, we introduce an effective congestion control scheme for TCP over hybrid wireless/wired networks comprising a multihop wireless IEEE 802.11 network and the wired Internet. We propose an adaptive pacing scheme at the Internet gateway for wired-to-wireless TCP flows. Furthermore, we analyze the causes for the unfairness of oncoming TCP flows and propose a scheme to throttle aggressive wired-to-wireless TCP flows at the Internet gateway to achieve nearly optimal fairness. Thus, we denote the introduced congestion control scheme TCP with Gateway Adaptive Pacing (TCP-GAP). For wireless-to-wired flows, we propose an adaptive pacing scheme at the TCP sender. In contrast to previous work, TCP-GAP does not impose any control traffic overhead for achieving fairness among active TCP flows. Moreover, TCP-GAP can be incrementally deployed because it does not require any modifications of TCP in the wired part of the network and is fully TCP-compatible. Extensive simulations using ns-2 show that TCPGAP is highly responsive to varying traffic conditions, provides nearly optimal fairness in all scenarios and achieves up to 42% more goodput than TCP NewReno.

This paper introduces an effective congestion control pacing scheme for TCP over multihop wireless networks with Internet connectivity. The pacing scheme is implemented at the wireless TCP sender as well as at the Internet gateway, and reacts according to the direction of TCP flows running across the wireless network and the Internet. Moreover, we analyze the causes for the unfairness of oncoming TCP flows and propose a scheme to throttle aggressive wired-to-wireless TCP flows at the Internet gateway to achieve nearly optimal fairness. The proposed scheme, which we denote as TCP with Gateway Adaptive Pacing (TCP-GAP), does not impose any control traffic overhead for achieving fairness among active TCP flows and can be incrementally deployed since it does not require any modifications of TCP in the wired part of the network. In an extensive set of experiments using ns-2 we show that TCP-GAP is highly responsive to varying traffic conditions, provides nearly optimal fairness in all scenarios and achieves up to 42% more goodput for FTP-like traffic as well as up to 70% more goodput for HTTP-like traffic than TCP NewReno. We also investigate the sensitivity of the considered TCP variants to different bandwidths of the wired and wireless links with respect to both aggregate goodput and fairness.

碩士
國立臺灣大學
電機工程學研究所
93
Due to the fast growth of high bandwidth network, real-time multimedia applications become increasingly popular. Real-time multimedia applications do not use Transmission Control Protocol (TCP) but adopt User Datagram Protocol (UDP) as transport mechanism, which may lead unfair bandwidth allocation or even shut down TCP traffic. In this thesis, a new TCP-friendly congestion control algorithm is proposed to ensure coexistence with TCP traffic, and better qualify in real-time multimedia application performance. The new algorithm is called the Hybrid TCP-Friendly Congestion Avoidance Control (HTCAC). Different from traditional TCP-friendly control algorithms, HTCAC is an active control system which does not passively wait for the happening of congestion but use Round Trip Time (RTT) information to adjust network transmission rate and to avoid network congestion!

碩士
國立彰化師範大學
資訊工程學系
100
Transmission Control Protocol (TCP) is a widely used end-to-end transport protocol in the Internet. It is given the task to protect the Internet from collapse as well as to make a good use of network resources. However, the congestion control of standard TCP (Reno) functions poorly in high speed networks because of its slow response with large congestion windows. Therefore, standard TCP may become the performance bottleneck as the bandwidth of Internet continues to grow. In this paper, we propose a new variant of hybrid TCP, Concave Convex TCP (CCTCP), to overcome this issue. CCTCP revises Reno’s congestion avoidance phase by appending delay-based features. It dynamically switches its state between concave and convex states depending on historical records and network conditions to adjust its window size. Through ns-2 based simulations and experiments on the Linux platform we find that, as compared to other hybrid TCPs, CCTCP can achieve a high performance with minor dependence on bottleneck buffer sizes under a variety of network bandwidths, especially in high speed networks. Moreover, it coexists with standard TCP by higher degrees of TCP-friendliness.

碩士
朝陽科技大學
資訊工程系碩士班
96
Rapid advances in wireless networks and mobile communications achieve ubiquitous access to the plentiful resources in the Internet and construct an all IP-based environment, in which most IP data packets transmitted through the TCP connections. A typical all IP network may consist of different wireless networks, e.g., the IEEE 802.11 WLANs and 3G/HSDPA/HSUPA cellular systems, and then form a heterogeneous wireless network. In TCP/IP transmissions, the TCP congestion control operates well in the wired network, but it is difficult to determine an accurate congestion window in a heterogeneous wireless network. The primary reason is that TCP connections are affected not only by networks congestion but also by wireless error links. Thus, this paper proposes a cross layer-based adaptive window congestion control, namely Cross Layer Logarithmic Increase Adaptive Decrease, CL-LIAD, for TCP congestion control in the heterogeneous wireless networks. CL-LIAD deploys three significant contributions. First, we propose a novel cross layer mechanism that enables the receiver’s TCP protocol to carry the MAC-layer wireless state information to the sender through the ACK option. Second, an adaptive bandwidth expectation algorithm is proposed to predict available bandwidth, and thus accurately determine the congestion window. Third, in the Congestion Avoidance (CA) phase, we propose a Logarithmic Increase algorithm to increase cwnd while receiving each ACK after causing three duplicate ACKs. In addition, we analyze the congestion window and throughput under different packet loss rate by determining a closed-form expression. Furthermore, the state transition diagram of CL-LIAD is detailed. Numerical results demonstrate that the proposed CL-LIAD outperforms other approaches in goodput, fairness, and friendliness under diverse topologies of the heterogeneous wireless network. Especially, in the case of 10% packet loss rate in wireless links, the proposed approach increases goodput up to 111% and 225% as compared with LogWestwood+ and NewReno, respectively.

碩士
國立中央大學
資訊工程研究所
95
In wireless sensor networks, the congestion occurs when offered traffic load exceeds available capacity of sensor nodes. In most applications, every sensor node will send its sensing event to sink node and result in the sensors closer to the sink experiencing congestion. Congestion may cause packets loss, lower network throughput and waste energy of sensors. To address this challenge, we propose a distributed algorithm that mitigates congestion and allocates appropriate source rate to sink node for sensor networks. The proposed algorithm is a hybrid congestion control protocol which is considered not only the packets delivery rate but remaining buffer size of each node. Our protocol can avoid packets drop due to traffic congestion and improve the network throughput. The simulation results show that the performance of our protocol is better than the previous works.

We introduce an adaptive pacing scheme to overcome the drawbacks of TCP in wireless mesh networks with Internet connectivity. The pacing scheme is implemented at the wireless TCP sender as well as at the mesh gateway, and reacts according to the direction of TCP flows running across the wireless network and the Internet. TCP packets are transmitted rate-based within the TCP congestion window according to the current out-of-interference delay and the coefficient of variation of recently measured round-trip times. Opposed to the majority of previous work which builds on simulations, we implement a Linux prototype of our approach and evaluate its feasibility in a real 20-node mesh testbed. In an experimental performance study, we compare the goodput and fairness of our approach against the widely deployed TCP NewReno. Experiments show that our approach, which we denote as Mesh Adaptive Pacing (MAP), can achieve up to 150% more goodput than TCP NewReno and significantly improves fairness between competing flows. MAP is incrementally deployable since it is TCP-compatible, does not require cross-layer information from intermediate nodes along the path, and requires no modifications in the wired domain.

by Fu Chengpeng.
Thesis (Ph.D.)--Chinese University of Hong Kong, 2001.
Includes bibliographical references (p. 102-119).
Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web.
Mode of access: World Wide Web.
Abstracts in English and Chinese.

Sharing the resources of multiple wireless networks with overlapped coverage areas has a potential of improving the transmission throughput. However, in the existing frameworks, the improvement cannot be achieved in congestion scenarios because of independent congestion control procedures among the end-to-end paths. Although various network characteristics make the congestion control complex, this variety can be useful in congestion avoidance if the networks cooperate with each other. When congestion happens in an end-to-end path, it is inevitable to have a packet transmission rate less than the minimum requested rate due to congestion window size adjustments. Cooperation among networks can help to avoid this problem for better service quality. When congestion is predicted for one path, some of the on-going packets can be sent over other paths instead of the congested path. In this way, the traffic can be shifted from a congested network to others, and the overall transmission throughput does not degrade in a congestion scenario. However, cooperation is not always advantageous since the throughput of cooperative transmission in an uncongested scenario can be less than that of non-cooperative transmission due to cooperation costs such as cooperation setup time, additional signalling for cooperation, and out-of-order packet reception. In other words, a trade-off exists between congestion avoidance and cooperation cost. Thus, cooperation should be triggered only when it is beneficial according to congestion level measurements. In this research, our aim is to develop an efficient cooperative congestion control scheme for a heterogeneous wireless environment. To this end, a cooperative congestion control algorithm is proposed, in which the state of an end-to-end path is provided at the destination terminal by measuring the queuing delay and estimating the congestion level. The decision on when to start/stop cooperation is made based on the network characteristics, instantaneous traffic condition, and the requested quality of service (QoS). Simulation results demonstrate the throughput improvement of the proposed scheme over non-cooperative congestion control.

碩士
國立政治大學
資訊科學學系
95
Communication networks have evolved tremendously in the past decades. TCP is the most dominant and deployed end-to-end transport protocol across Internet today and will continue to be in the foresee future. It has numerous enhancing versions for wired network such as TCP Reno, TCP NewReno and TCP Vegas to improve the drawbacks of initial version of TCP. As IEEE 802.11 wireless network technology gains popularity, TCP is very likely to be popular for existing applications so far. However due to unawareness of network conditions, regular TCP is not able to fully control the limited resources and distinguish packet loss from congestion loss and random loss. Based on such implicit assumption, many studies have shown this would results in serious performance degradation in wireless environment. In this paper, we proposed a new TCP congestion control mechanism by router-assisted approach which is inspired by the concept of each wireless node playing the roles of terminal and router simultaneously. Based on the information feedback from routers, sender is able to adjust the sending speed dynamically in order to avoid overshooting problem. We also proposed a multilevel date rate adjustment method to control the date rate more precisely. Finally we evaluate the performance of our approach by NS2 simulator. Our proposed protocol has 5~10% higher throughput than TCP NewReno and much less number of retransmission. The fairness requirement is also achieved while our proposed protocol coexists with other major TCP variants.

In this paper, we introduce a novel congestion control algorithm for TCP over multihop IEEE 802.11 wireless networks implementing rate-based scheduling of transmissions within the TCP congestion window. We show how a TCP sender can adapt its transmission rate close to the optimum using an estimate of the current 4-hop propagation delay and the coefficient of variation of recently measured round-trip times. The novel TCP variant is denoted as TCP with Adaptive Pacing (TCP-AP). Opposed to previous proposals for improving TCP over multihop IEEE 802.11 networks, TCP-AP retains the end-to-end semantics of TCP and does neither rely on modifications on the routing or the link layer nor requires cross-layer information from intermediate nodes along the path. A comprehensive simulation study using ns-2 shows that TCP-AP achieves up to 84% more goodput than TCP NewReno, provides excellent fairness in almost all scenarios, and is highly responsive to changing traffic conditions.

碩士
國立彰化師範大學
資訊工程學系
96
Neither the current TCP protocol nor the standard backoff algorithm of IEEE 802.11 protocol is able to distinguish corruption loss from congestion or collision loss. Hence, high transmission errors and a varying latency inherent in wireless channel would have a seriously adverse effect on the performance of TCP. In this paper, we propose a novel and pragmatic cross-layer approach with joint congestion and contention window control scheme to improve the performance of TCP in IEEE 802.11 wireless environments. In addition to theoretical analysis, simulations are conducted to evaluate the proposed scheme. As it turns out, our design indeed provides a more efficient solution for frequent transmission loss and enables TCP to distinguish between congestion loses and transmission errors, thus to take proper remedial actions.

In this paper, we introduce a clean-slate architecture for improving the delivery of data packets in IEEE 802.11 wireless mesh networks. Opposed to the rigid TCP/IP layer architecture which exhibits serious deficiencies in such networks, we propose a unitary layer approach that combines both routing and transport functionalities in a single layer. The new Mesh Transmission Layer (MTL) incorporates cross-interacting routing and transport modules for a reliable data delivery based on the loss probabilities of wireless links. Due to the significant drawbacks of standard TCP over IEEE 802.11, we particularly focus on the transport module, proposing a pure rate-based approach for transmitting data packets according to the current contention in the network. By considering the IEEE 802.11 spatial reuse constraint and employing a novel acknowledgment scheme, the new transport module improves both goodput and fairness in wireless mesh networks. In a comparative performance study, we show that MTL achieves up to 48% more goodput and up to 100% less packet drops than TCP/IP, while maintaining excellent fairness results.

This dissertation is a study on the design and analysis of novel, optimal routing and rate control algorithms in wireless, mobile communication networks. Congestion control and routing algorithms upto now have been designed and optimized for wired or wireless mesh networks. In those networks, optimal algorithms (optimal in the sense that either the throughput is maximized or delay is minimized, or the network operation cost is minimized) can be engineered based on the classic time scale decomposition assumption that the dynamics of the network are either fast enough so that these algorithms essentially see the average or slow enough that any changes can be tracked to allow the algorithms to adapt over time. However, as technological advancements enable integration of ever more mobile nodes into communication networks, any rate control or routing algorithms based, for example, on averaging out the capacity of the wireless mobile link or tracking the instantaneous capacity will perform poorly. The common element in our solution to engineering efficient routing and rate control algorithms for mobile wireless networks is to make the wireless mobile links seem as if they are wired or wireless links to all but few nodes that directly see the mobile links (either the mobiles or nodes that can transmit to or receive from the mobiles) through an appropriate use of queuing structures at these selected nodes. This approach allows us to design end-to-end rate control or routing algorithms for wireless mobile networks so that neither averaging nor instantaneous tracking is necessary, as we have done in the following three networks. A network where we can easily demonstrate the poor performance of a rate control algorithm based on either averaging or tracking is a simple wireless downlink network where a mobile node moves but stays within the coverage cell of a single base station. In such a scenario, the time scale of the variations of the quality of the wireless channel between the mobile user and the base station can be such that the TCP-like congestion control algorithm at the source can not track the variation and is therefore unable to adjust the instantaneous coding rate at which the data stream can be encoded, i.e., the channel variation time scale is matched to the TCP round trip time scale. On the other hand, setting the coding rate for the average case will still result in low throughput due to the high sensitivity of the TCP rate control algorithm to packet loss and the fact that below average channel conditions occur frequently. In this dissertation, we will propose modifications to the TCP congestion control algorithm for this simple wireless mobile downlink network that will improve the throughput without the need for any tracking of the wireless channel. Intermittently connected network (ICN) is another network where the classic assumption of time scale decomposition is no longer relevant. An intermittently connected network is composed of multiple clusters of nodes that are geographically separated. Each cluster is connected wirelessly internally, but inter-cluster communication between two nodes in different clusters must rely on mobile carrier nodes to transport data between clusters. For instance, a mobile would make contact with a cluster and pick up data from that cluster, then move to a different cluster and drop off data into the second cluster. On contact, a large amount of data can be transferred between a cluster and a mobile, but the time duration between successive mobile-cluster contacts can be relatively long. In this network, an inter-cluster rate controller based on instantaneously tracking the mobile-cluster contacts can lead to under utilization of the network resources; if it is based on using long term average achievable rate of the mobile-cluster contacts, this can lead to large buffer requirements within the clusters. We will design and analyze throughput optimal routing and rate control algorithm for ICNs with minimum delay based on a back-pressure algorithm that is neither based on averaging out or tracking the contacts. The last type of network we study is networks with stationary nodes that are far apart from each other that rely on mobile nodes to communicate with each other. Each mobile transport node can be on one of several fixed routes, and these mobiles drop off or pick up data to and from the stationaries that are on that route. Each route has an associated cost that much be paid by the mobiles to be on (a longer route would have larger cost since it would require the mobile to expend more fuel) and stationaries pay different costs to have a packet picked up by the mobiles on different routes. The challenge in this type of network is to design a distributed route selection algorithm for the mobiles and for the stationaries to stabilize the network and minimize the total network operation cost. The sum cost minimization algorithm based on average source rates and mobility movement pattern would require global knowledge of the rates and movement pattern available at all stationaries and mobiles, rendering such algorithm centralized and weak in the presence of network disruptions. Algorithms based on instantaneous contact, on the contrary, would make them impractical as the mobile-stationary contacts are extremely short and infrequent.
text