An effective feedback control mechanism for DiffServ architecture

The differentiated services (DiffServ) is proposed to provide packet level service differentiations in a scalable manner. To provide an end-to-end service differentiation to users having a connection over multiple domains, as well as a flow marker, an intermediate marker is necessary at the edge routers, and it should not be operated at a flow level due to a scalability problem. Due to this operation requirement, the intermediate marker has a fairness problem among the transmission control protocol (TCP) flows since TCP flows have intrinsically unfair throughputs due to the TCP's congestion control algorithm. Moreover, it is very difficult to resolve this problem without individual flow state information such as round trip time (RTT) and sending rate of each flow. In this paper, to resolve this TCP fairness problem of an intermediate marker, we propose a fair scalable marker (FSM) as an intermediate marker which works with a source flow three color marker (sf-TCM) operating as a host source marker. The proposed fair scalable marker improves the fairness among the TCP flows with different RTTs without per-flow management. Through the simulations, we show that the FSM can improve TCP fairness as well as link utilization in multiple domain DiffServ networks.

Classical Transmission Control Protocol (TCP) designs have never considered the identity of the competing transport protocol as useful information to TCP sources in congestion control mechanisms. When competing against a TCP flow on a bottleneck link, a User Datagram Protocol (UDP) flow can unfairly occupy the entire link bandwidth and suffocate all TCP flows on the link. If it were possible for a TCP source to know the type of transport protocol that deprives it of link access, perhaps it would be better for the TCP source to react in a way which prevents total starvation. In this paper, we use coefficient of variation and power spectral density of throughput traces to identify the presence of UDP transport protocols that compete against TCP flows on bottleneck links. Our results show clear traits that differentiate the presence of competing UDP flows from TCP flows independent of round-trip times variations. Signatures that we identified include an increase in coefficient of variation whenever a competing UDP flow joins the bottleneck link for the first time, noisy spectral density representation of a TCP flow when competing against a UDP flow in the bottleneck link, and a dominant frequency with outstanding power in the presence of TCP competition only. In addition, the results show that signatures for congestion caused by competing UDP flows are different from signatures due to congestion caused by competing TCP flows regardless of their round-trip times. The results in this paper present the first steps towards development of more ’intelligent’ congestion control algorithms with added capability of knowing the identity of aggressor protocols against TCP, and subsequently using this additional information for rate control.

Many smart healthcare centers are deploying long distance, high bandwidth networks in their computer network infrastructure and operation. Transmission control protocol (TCP) is responsible for reliable and secure communication of data in these medial infrastructure networks. TCP is reliable and secure due to its congestion control mechanism, which is responsible for detecting and reacting to the congestion in the network. Many TCP congestion control mechanisms have been developed previously for different operating systems. TCP CUBIC, TCP Compound, and TCP Fusion are the default congestion control mechanism in Linux, Microsoft Windows, and Sun Solaris operating systems, respectively. The earliest congestion control mechanism Standard TCP acts as the trademark congestion control mechanism. The exponential growth of congestion window ( $cwnd$ ) in slow start phase of the TCP CUBIC causes burst losses of packets, and TCP flows did not share available link bandwidth fairly. The prime aim of this paper is to enhance the performance of TCP CUBIC for long distance, high bandwidth secured networks to achieve better performance in medical infrastructure, concerning packet loss rate, protocol fairness, and convergence time. In this paper, congestion control module for slow start is proposed, which reduces the effect of the exponential growth of $cwnd$ by designing the new limits of $cwnd$ size in slow start phase, which in turn decreases the packet loss rate in healthcare networks. NS-2 is used to simulate the experiments of enhanced TCP CUBIC and state-of-the-art congestion control mechanisms. Results show that the performance of enhanced TCP CUBIC outperforms by 18% as compared with the state-of-the-art congestion control mechanisms.

The transmission control protocol (TCP) is one of the most important Internet protocols. It provides reliable transport services between two end-hosts. Since TCP performance affects overall network performance, many studies have been done to model TCP performance in the steady state. However, recent researches have shown that most TCP flows are short-lived. Therefore, it is more meaningful to model TCP performance in relation to the initial stage of short-lived flows. In addition, the next-generation Internet will be an unified all-IP network that includes both wireless and wired networks integrated together. In short, modelling short-lived TCP flows in wireless networks constitutes an important axis of research. In this paper, we propose simple wireless TCP models for short-lived flows that extend the existing analytical model proposed in [IEEE Commun. Lett. 2002; 6(2):85–88]. In terms of wireless TCP, we categorized wireless TCP schemes into three types: end-to-end scheme, split connection scheme, and local retransmission scheme, which is similar to the classification proposed in [IEEE/ACM Trans. Networking 1997; 756–769]. To validate the proposed models, we performed ns-2 simulations. The average differences between the session completion time calculated using the proposed model and the simulation result for three schemes are less than 9, 16, and 7 ms, respectively. Consequently, the proposed model provides a satisfactory means of modelling the TCP performance of short-lived wireless TCP flows. Copyright © 2005 John Wiley & Sons, Ltd.

In Internet gateways with active queue management, effective algorithms are often used to enhance the traffic flows through the gateway. The random early detection (RED) algorithm is a widely used algorithm for controlling the transmission control protocol (TCP) flows. Specifically, the RED algorithm has deliberately introduced randomness into the explicit congestion notification (ECN) marking mechanism in order to enhance the fairness and to avoid synchronization of TCP flows through the gateway. As randomness plays a key role in providing better performance for RED gateways than the DropTail gateways, knowledge of the effect of the random mechanism is important to the understanding the behavior of the system. The fluid flow model (FFM), ignoring the randomness of ECN markings, has provided a fast alternative to the slow ns-2 network simulation tool in studying the TCP flow in the RED gateway. In this brief, the random ECN marking mechanism in RED gateways has been studied. The randomness of the RED ECN marking algorithm is implemented into the FFM. Verified using ns-2 simulations, the proposed model shows better dynamical performance than FFM in both the single-bottleneck dumbbell network and in networks with interactive bottlenecks.

Analytical modeling of TCP flow in wireless LANs

The differentiated services (DiffServ) architecture for provisioning quality of service (QoS) in the Internet provides a flexible framework for supporting a variety of services in heterogeneous environments. The assured forwarding (AF) per hop behaviour is a means of providing differential treatment of various DiffServ classes through achieving higher forwarding probabilities for higher priority classes. Extending the AF model to support fine grain QoS specification in W-CDMA environments allows mobile users in third-generation systems to benefit from the economical savings made possible by resource sharing among different classes of aggregated traffic. We consider extending the AF model to support the delivery of quasi constant bit-rate (QCBR) traffic streams on the downlink. In our study, each QCBR stream is assumed to have a prescribed average bit-rate and time duration, and is expected to be transmitted to completion without interruption at the requested bit-rate. The availability of such a service is beneficial for transmitting real-time traffic to mobile devices with limited power and buffering resources. The proposed mechanisms are based on integrating a suitable power-sharing structure for achieving differential treatment between classes with a crude power prediction algorithm for estimating the probability that no forced termination will occur at certain instants in the future. We compare the performance of a non-predictive scheme with a predictive scheme for different AF classes and source information bit-rates. Our results show that significant improvement in the forwarding probability, throughput, and power utilization for the DiffServ classes can be attained using the predictive scheme.

Due to the development and evolution in Cognitive Radio Networks (CRNs), the performance of Transmission Control Protocol (TCP) over CRN has become a crucial factor. TCP ensures the reliable end-to-end delivery of data, which is very important for the quality of service (QoS) requirements. TCP performance is severely affected by channel switching and channel availability in CRNs. The frequent occurrence of such events is due to CRNs nature which unnecessarily reduces the transmission rate which results in degrading the overall performance. In this paper, the authors investigate the limitations of using TCP NewReno (i.e., standard TCP) over CRNs, and further propose a cross-layer-design protocol based on an interaction between transport layer and lower layers to improve TCP performance over CRNs. The design is called Cross-Layer Design for TCP (TCP-CLD). It primarily aims at improving the TCP performance over CRNs via controlling and adapting a few selected TCP parameters. Specifically, the authors propose freezing the data transmission and re-transmission timeout (RTO) of TCP during the time of channel switching and channel unavailability. In addition, the authors update the size of congestion window (cwnd) of TCP based on the bandwidth and buffer size of relay nodes. Moreover, in the case of channel switching, the authors update the Round Trip Time (RTT) of TCP based on the new route delay. Simulation experiments have been conducted using ns-2 network simulator tool whereas results show that the proposed design achieves a significant performance improvement compared with the standard TCP NewReno protocol and the other known protocol, namely, TCP-CRAHN.

To quantitatively measure quality of service (QoS) in Network on Chip (NoC), several related aspects of the network service are often considered, such as end to end delay (EED), Throughput (Thp), Packet loss rate (PLR), etc. However, until now, no standard method of performance measurement and fewer techniques have been used to provide its definition. In fact, few papers have developed different methods to modelize QoS in NoC and provided an efficient and flexible way to monitor QoS. The originality of our approach is based on a proposition of a QoS–intellectual property module in NoC architecture to improve network performances. We implement an approach of QoS metrics modeling for NoC, using Analytic Hierarchy Process (AHP) on multi-parameter and multi-application for 4×4 mesh NoC environment. The results have shown that our QoS modeling approach is proven successful in providing a quantifiable representation. Therefore, QoS arbiter module interacts with other routers, the link utilization is balanced and network performance improved.

<p>The demand for internet applications has increased rapidly. Providing quality of service (QoS) requirements for varied internet application is a challenging task. One important factor that is significantly affected on the QoS service is the transport layer. The transport layer provides end-to-end data transmission across a network. Currently, the most common transport protocols used by internet application are TCP (Transmission Control Protocol) and UDP (User Datagram Protocol). Also, there are recent transport protocols such as DCCP (data congestion control protocol), SCTP (stream congestion transmission protocol), and TFRC (TCP-friendly rate control), which are in the standardization process of Internet Engineering Task Force (IETF). In this paper, we evaluate the performance of UDP, DCCP, SCTP and TFRC protocols for different traffic flows: data transmission, video traffic, and VOIP in wired networks. The performance criteria used for this evaluation include throughput, end to end delay, and packet loss rate. Well-known network simulator NS-2 used to implement the UDP, DCCP, SCTP, and TFRC protocols performance comparison. Based on the simulation results, the performance throughput of SCTP and TFRC is better than UDP. Moreover, DCCP performance is superior SCTP and TFRC in term of end-to-end delay.</p>

Power Line Communications (PLC) is regarded as the promising home networking technology for smart grid application since the PLC technology enables data transmission via the existing power line. The cooperation between PLC and Wireless Local Area Network (WLAN) will be an effective hybrid solution for Home Area Networks (HANs). In this paper, we consider that Transmission Control Protocol (TCP) packets are transmitting in heterogenous communications network based on PLC and WLAN. The existing TCP congestion control has low effectiveness due to the error-prone power line channel. Therefore, we propose an adaptive rate control to improve TCP performance in the hybrid PLC/WLAN HANs. We exploit the communications mechanism of WLAN to improve the PLC link utilization. The methodology of Heuristic Dynamic Programming (HDP) is employed to obtain the optimal rate policy. The bottleneck link capacity is estimated by neural network training. The sending rate is adapted according to the long-term quality of service (QoS) of applications. Simulation results illustrate that the proposed rate control effectively filters out the negative impact from the instantaneous impulse noise in PLC channel and outperforms the existing TCP.

This paper studies an important problem in the IEEE 802.11 distributed coordination function (DCF)-based wireless local area network (WLAN): how well can the network support quality of service (QoS). Specifically, this paper analyzes the network's performance in terms of maximum protocol capacity or throughput, delay, and packet loss rate. Although the performance of the 802.11 protocol, such as throughput or delay, has been extensively studied in the saturated case, it is demonstrated that maximum protocol capacity can only be achieved in the nonsaturated case and is almost independent of the number of active nodes. By analyzing packet delay, consisting of medium access control (MAC) service time and waiting time, accurate estimates were derived for delay and delay variation when the throughput increases from zero to the maximum value. Packet loss rate is also given for the nonsaturated case. Furthermore, it is shown that the channel busyness ratio provides precise and robust information about the current network status, which can be utilized to facilitate QoS provisioning. The authors have conducted a comprehensive simulation study to verify their analytical results and to tune the 802.11 to work at the optimal point with maximum throughput and low delay and packet loss rate. The simulation results show that by controlling the total traffic rate, the original 802.11 protocol can support strict QoS requirements, such as those required by voice over Internet protocol (VoIP) or streaming video, and at the same time achieve high channel utilization.

Transmission control protocol (TCP) provides reliable connection-oriented services between any two end systems on the Internet. With TCP congestion control algorithm, multiple TCP connections can share network and link resources simultaneously. These TCP congestion control mechanisms have been operating effectively in wired networks. However, performance of TCP connections degrades rapidly in wireless and lossy networks. To sustain the throughput performance of TCP connections in wireless networks, design modifications may be required accordingly in the TCP flow control algorithm, and potentially, in association with other protocols in other layers for proper adaptations. In this chapter, we explain the limitations of the latest TCP congestion control algorithm, and then review some popular designs for TCP connections to operate effectively in wireless mesh network infrastructure.

Scalable Quality of Service (QoS) control is of paramount importance to effectively enable a seamless convergence of the rapidly evolving Peer-to-Peer (P2P) overlay communications over the Internet since the latter only supports best-effort service paradigm. For example, the European Union (EU) funded ROMEO project is focusing on a joint use of DVB-T2 and P2P overlay networks for live multimedia content sharing and collaboration among multiple users. This raises a strong need that the media packets transmitted through the P2P overlay delivery system must arrive earlier enough at the end users to assure a proper synchronization of the multiple views that may be received via the hybrid network. For this purpose, the P2P network must assure a certain QoS guarantee in terms of bandwidth, delay, jitter, and loss. More importantly, the control must be scalable to prevent excessive signalling and the related processing overhead, usually suffered in the traditional per-flow QoS control approaches. In this view, recent research effort claimed that the Internet resources can be efficiently over-provisioned (booking more resources in-advance) in such a way to allow differentiation of QoS control with reduced signalling overhead and increased resource utilization. This approach, however, needs further investigations for proper integration into innovative networking architectural designs to achieve performance. In addition to that, and in order to provide an end-to-end QoS, a mechanism to enforce prioritization policies within the customer’s access network is also needed. Hence, this chapter proposes a cross-layer control architecture that takes advantage of the Internet resources over-provisioning and the QoS policy enforcement within access networks to facilitate rapid development of P2P applications. The design aims to alleviate the requirements of buffers and the need for adaptation on end users’ devices, thus allowing for cost-effective and rapid development of attractive services in similar hybrid content delivery networks.

QNoC: QoS architecture and design process for network on chip