Multiple TCP Flows Research Articles

MPTCP Tunnel: An Architecture for Aggregating Bandwidth of Heterogeneous Access Networks

Fixed and cellular networks are two typical access networks provided by operators. Fixed access network is widely employed; nevertheless, its bandwidth is sometimes not sufficient enough to meet user bandwidth requirements. Meanwhile, cellular access network owns unique advantages of wider coverage, faster increasing link speed, more flexible deployment, and so forth. Therefore, it is attractive for operators to mitigate the bandwidth shortage by bundling these two. Actually, there have been existing schemes proposed to aggregate the bandwidth of two access networks, whereas they all have their own problems, like packet reordering or extra latency overhead. To address this problem, we design new architecture, MPTCP Tunnel, to aggregate the bandwidth of multiple heterogeneous access networks from the perspective of operators. MPTCP Tunnel uses MPTCP, which solves the reordering problem essentially, to bundle multiple access networks. Besides, MPTCP Tunnel sets up only one MPTCP connection at play which adapts itself to multiple traffic types and TCP flows. Furthermore, MPTCP Tunnel forwards intact IP packets through access networks, maintaining the end‐to‐end TCP semantics. Experimental results manifest that MPTCP Tunnel can efficiently aggregate the bandwidth of multiple access networks and is more adaptable to the increasing heterogeneity of access networks than existing mechanisms.

Frame Transmission Efficiency-Based Cross-Layer Congestion Notification Scheme in Wireless Ad Hoc Networks.

Different from the traditional wired network, the fundamental cause of transmission congestion in wireless ad hoc networks is medium contention. How to utilize the congestion state from the MAC (Media Access Control) layer to adjust the transmission rate is core work for transport protocol design. However, recent works have shown that the existing cross-layer congestion detection solutions are too complex to be deployed or not able to characterize the congestion accurately. We first propose a new congestion metric called frame transmission efficiency (i.e., the ratio of successful transmission delay to the frame service delay), which describes the medium contention in a fast and accurate manner. We further present the design and implementation of RECN (ECN and the ratio of successful transmission delay to the frame service delay in the MAC layer, namely, the frame transmission efficiency), a general supporting scheme that adjusts the transport sending rate through a standard ECN (Explicit Congestion Notification) signaling method. Our method can be deployed on commodity switches with small firmware updates, while making no modification on end hosts. We integrate RECN transparently (i.e., without modification) with TCP on NS2 simulation. The experimental results show that RECN remarkably improves network goodput across multiple concurrent TCP flows.

Analysis of multiple flows using different high speed TCP protocols on a general network

We develop analytical tools for performance analysis of multiple TCP flows (which could be using TCP CUBIC, TCP Compound, TCP New Reno) passing through a multi-hop network. We first compute average window size for a single TCP connection (using CUBIC or Compound TCP) under random losses. We then consider two techniques to compute steady state throughput for different TCP flows in a multi-hop network. In the first technique, we approximate the queues as M/G/1 queues. In the second technique, we use an optimization program whose solution approximates the steady state throughput of the different flows. Our results match well with ns2 simulations.

FMTCP: A Fountain Code-Based Multipath Transmission Control Protocol

Ideally, the throughput of a Multipath TCP (MPTCP) connection should be as high as that of multiple disjoint single-path TCP flows. In reality, the throughput of MPTCP is far lower than expected. In this paper, we conduct an extensive simulation-based study on this phenomenon, and the results indicate that a subflow experiencing high delay and loss severely affects the performance of other subflows, thus becoming the bottleneck of the MPTCP connection and significantly degrading the aggregate goodput. To tackle this problem, we propose Fountain code-based Multipath TCP (FMTCP), which effectively mitigates the negative impact of the heterogeneity of different paths. FMTCP takes advantage of the random nature of the fountain code to flexibly transmit encoded symbols from the same or different data blocks over different subflows. Moreover, we design a data allocation algorithm based on the expected packet arriving time and decoding demand to coordinate the transmissions of different subflows. Quantitative analyses are provided to show the benefit of FMTCP. We also evaluate the performance of FMTCP through ns-2 simulations and demonstrate that FMTCP outperforms IETF-MPTCP, a typical MPTCP approach, when the paths have diverse loss and delay in terms of higher total goodput, lower delay, and jitter. In addition, FMTCP achieves high stability under abrupt changes of path quality.

Migrate to IPv6 on 6PE with MPLS-TE: Questions on Feasibility and Reluctance in Deployment?

Due to the exhaustion of IPv4 addresses, the transition to IPv6 has already begun using various transition mechanisms. In this paper comparison between two transition mechanism 6PE and 6to4 tunneling over MPLS (Multiprotocol Label Switching) based network has been performed. The experimental platform used is based on GNS3/Dynagen using emulated routers on which multiple TCP and UDP flows were generated between separated IPv6 network traversing through IPv4 based core. Traffic were generated using both transition mechanism and are compared to native IPv4 based MPLS network. Performance metric used are end-to-end delay, jitter and throughput, these parameters are also analyzed statistically using ANOVA, F-Test and T-Test to support a conclusion. Further, MPLS Traffic-Engineering (TE) using RSVP-TE signalling is applied, describing the implementation and performance test and analysis of TE tunnels on 6PE for IPv6 traffic. The same test-bed is used to implement and observe MPLS-TE behavior and throughput of 6PE based MPLS-TE to that of native IPv4 based MPLS-TE. Concluding, throughput of 6PE and 6to4 is around 99% and 96%, respectively, as compared to that of native IPv4.

Link Layer Correction Techniques and Impact on TCP’s Performance in IEEE 802.11 Wireless Networks

TCP performance degrades when end-to-end connections extend over wireless links which are characterized by high Bit Error Rate and intermittent connectivity. Such degradation is mainly accounted for TCP’s unnecessary congestion control actions while attempting TCP loss recovery. Several independent link loss recovery approaches are proposed by researchers to reduce number of losses visible at TCP. In this paper we first presented a survey of loss mitigation techniques at wireless link layer. Secondly performance evaluation for TCP through Type 0 Automatic Retransmission Request mechanism in erroneous Wireless LAN is presented. In particular, simulations are performed taking into account the wireless errors introduced over IEEE 802.11 link using a well-established 2-State Markov model. TCP performance is evaluated under different settings for maximum link retransmissions allowed for each frame. Simulation results show that, link retransmission improves TCP performance by reducing losses perceived at TCP sender. However, such improvement is often associated with adverse effect on other TCP parameters that may cost a lot in return under extreme network conditions. In this paper an attempt is made to observe impact of link retransmissions on the performance of multiple TCP flows competing with each other. The analysis presented in this paper signifies the scope for maximizing TCP’s throughput at the least possible cost.

SCTP Tunneling: Flow Aggregation and Burst Transmission to Save Energy for Multiple TCP Flows over a WLAN

SUMMARY To raise the energy efficiency of wireless clients, it is important to sleep in idle periods. When multiple network applications are running concurrently on a single wireless client, packets of each application are sent and received independently, but multiplexed at MAC-level. This uncoordinated behavior makes it difficult to control of sleep timing. In addition, frequent state transitions between active and sleep modes consume non-negligible energy. In this paper, we propose a transport-layer approach that resolves this problem and so reduces energy consumed by multiple TCP flows on a wireless LAN (WLAN) client. The proposed method, called SCTP tunneling, has two key features: flow aggregation and burst transmission. It aggregates multiple TCP flows into a single SCTP association between a wireless client and an access point to control packet transmission and reception timing. Furthermore, to improve the sleep efficiency, SCTP tunneling reduces the number of state transitions by handling multiple packets in a bursty fashion. In this study, we construct a mathematical model of the energy consumed by SCTP tunneling to assess its energy efficiency. Through numerical examples, we show that the proposed method can reduce energy consumption by up to 69%.

Analysis of TCP performance on multi-hop wireless networks: A cross layer approach

In Multi-Hop Wireless Networks (MHWNs), wireless nodes cooperate to forward traffic between end points that are not in direct communication range. Specifically, traffic is forwarded from a source towards its destination through intermediate nodes that form a wireless multi-hop chain. Researchers have studied the performance of TCP over chains discovering properties such as how the number of hops reduces chain throughput as neighboring links contend for the shared medium. Moreover, the presence of hidden terminals has also been shown to negatively affect performance of example chains. In this paper, we leverage recent characterization of how competing wireless links interact to develop an in-depth analysis of TCP performance over wireless chains. In particular, there are a number of possible modes of interference between competing links with distinct implications on performance and fairness; to our knowledge, this is the first work that studies the impact of these different modes on TCP chain performance. We classify chains according to interference modes considering both the forward (data) and reverse (acknowledgment) traffic. Chain geometry limits the types of chains that arise most frequently in practice. We evaluate TCP performance over the most frequently occurring chain types and observe significant performance differences between chains that have the same hop count. Different four-hop chains, for example, show a throughput difference of up to 25% and a retransmission overhead difference of over 90%. We discuss the implications of these differences on network performance: specifically, route instability and bandwidth usage generated. We extend this analysis to two single-hop TCP flows and quantify the effect of interference interactions between two flows. This study is a first step towards completely understanding the performance of multiple TCP flows over multiple hops in a MHWN.

Performance Analysis of Proactive and Reactive Transport Protocols using a Proactive Congestion Avoidance Model

paper attempts to establish the fact that Proactive transport protocols are better than conventional Reactive transport protocols in terms of overall utilization of resources. This has been proved analytically in this paper using a mathematical model developed for proactive congestion avoidance. The paper also provides a mathematical framework for Proactive Transport Protocol and predicts the congestion window evolution and throughput achieved by single and multiple proactive TCP flows. The congestion window evolution and throughput of a Proactive TCP connection are modeled as functions of the bandwidth, delay, packet size, number of concurrent TCP flows and the penalty factor used by the proactive protocols. Results obtained from the model are verified with ns2 simulation. The optimal value of the penalty factor is derived mathematically from the designed model. It has also been shown how the derived optimal values enhances the performance of the protocol and almost a near maximum utilization of resources is achieved by extensive simulation in ns2 considering both single and multiple TCP flows.

Network-supported TCP rate control for high-speed power line communications environments

High-speed power line communications (PLC) technology has great potential to serve as an infrastructure for home networking. To investigate the performance of end-to-end data flows over high definition (HD)-PLC, we have developed a new module on the NS-2 network simulator for HD-PLC that is driven by the particular patterns of the time-varying packet-error-rate as measured by experiments. To solve the problem on coexisting multiple TCP flows, a novel network-supported TCP rate control that appropriately modifies the advertized-window value in a TCP-ACK packet in response to the PLC network condition is presented and is shown to be effective through simulations.

An accurate closed-form formula for the throughput of long-lived TCP connections in IEEE 802.11 WLANs

There is a vast literature on the performance modeling of the 802.11 MAC protocol, but the interplay between the TCP dynamics and the self-regulating behavior of the 802.11 CSMA/CA access method has not been sufficiently investigated. In addition, it has been observed that TCP stations in a WLAN are sporadically active due to the TCP flow control mechanisms. This makes traditional saturated models of the network capacity of 802.11 WLANs not very accurate. We presented a rigorous analytical model to calculate the distribution of the number of active TCP stations in a WLAN with multiple long-lived TCP flows [R. Bruno, M. Conti, E. Gregori, Performance modelling and measurements of TCP transfer throughput in 802.11-based WLAN, in: Proceedings of the IEEE MsWiM 2006, Torremolinos, Malaga, Spain, 2006, pp. 4-11, [1]]. Starting from this analysis, in this paper we derive a simple but accurate closed-form expression of the per-connection TCP throughout as a function of the average duration of collisions, the average backoff period and the TCP packet size. We validate this formula through performance tests carried out on a real WLAN. Then, we use our model to mathematically study the optimal minimum contention window size to maximize the TCP throughput. Our analytical results indicate that the initial window size specified in the 802.11b standard is about two times larger than the optimal one. However, we also show that the TCP throughput performance is not very sensitive to changes of the minimum contention window size, and that the TCP flows significantly underutilize the channel bandwidth also in the optimal MAC operating state.

On the throughput and fairness performance of TCP over ethernet passive optical networks

The Ethernet Passive Optical Network (EPON) is a promising solution for metropolitan optical networks. In EPON, the downstream channel is a point-to-multipoint broadcast network and the upstream channel is a multipoint-to-point network based on a polling mechanism. In this paper, we model the performance of TCP over EPON with respect to the aggregate throughput and fairness. We calculate the maximum aggregate throughput of TCP flows on each of the upstream and downstream channels, and identify the fairness problem with existing polling schemes for EPON. We then derive two tight bounds of the fairness index for downloading users, and propose two polling mechanisms to improve unfair resource sharing among multiple TCP flows. The performance of the proposed mechanisms is evaluated via ns-2 simulations. The results show that our mechanisms can improve fairness among downloading TCP flows without degrading the aggregate throughput and fairness of uploading TCP flows.

TCP BaLDE for Improving TCP Performance over Heterogeneous Networks

Network congestion and random errors of wireless link are two well-known noteworthy parameters which degrade the TCP performance over heterogeneous networks. We put forward a novel end-to-end TCP congestion control mechanism, namely TCP BaLDE (Bandwidth and Loss Differentiation Estimate), in which the TCP congestion control categorizes the reason of the packet loss by estimating loss differentiation in order to control the packet transmission rate appropriately. While controlling transmission rate depends on the available bandwidth estimation which is apprehended by the bandwidth estimation algorithm when the sender receives a new ACK with incipient congestion signal, duplicates ACKs or is triggered by retransmission timeout event. Especially, this helps the sender to avoid router queue overflow by opportunely entering the congestion avoidance phase. In simulation, we experimented under numerous different network conditions. The results show that TCP BaLDE can achieve robustness in aspect of stability, accuracy and rapidity of the estimate in comparison with TCP Westwood, and tolerate ACK compression. It can achieve better performance than TCP Reno and TCP Westwood. Moreover, it is fair on bottleneck sharing to multiple TCP flows of the same TCP version, and friendly to existing TCP version.

The effect of packet reordering in a backbone link on application throughput

Packet reordering in the Internet is a well-known phenomenon. As the delay and speed of backbone links continue to increase, what used to be a negligible amount of packet reordering may now, combined with some level of dropped packets, cause multiple invocations of fast recovery within a TCP window. This may result in a significant drop in link utilization and hence in application throughput. What adds to the difficulty is that packet reordering is a silent problem. It may result in significant application throughput degradation while leaving little to no trace. In this article we try to measure and quantify the effect of reordering packets in a backbone link that multiplexes multiple TCP flows on application throughput. Different operating systems and delay values as well as various types of flow mixes were tested in a laboratory setup. The results show that only a small percentage of reordered packets, by at least three packet locations, in a backbone link can cause significant degradation of application throughput. Long flows are affected most. Due to the potential impact of this phenomenon, minimization of packet reordering as well as mitigating the effect algorithmically should be considered.

Investigating the performance of TCP in mobile ad hoc networks

Mobile ad hoc networks have several inherent characteristics (e.g. dynamic topology, time-varying and bandwidth constrained wireless channels, multi-hop routing, and distributed control and management). The goal of this work is to investigate the impact of these characteristics on the performance of TCP. First, we investigate throughput performance of TCP as a function of path length (i.e. multiple wireless hops), node mobility, and traffic intensity. Next, we examine the ‘fairness’ of the ad hoc network with regard to equal sharing of network bandwidth among multiple TCP flows. Third, we evaluate the impact of two on-demand routing protocols (i.e. AODV and DSR) on the throughput of TCP. Finally, a factorial design experiment is conducted to quantify the effects and interactions of three factors, which influence the throughput of TCP. These factors include routing, node speed, and node pause time. Two key results were observed. Results show that traffic intensity (e.g. number of concurrent flows) is significantly affects TCP throughput, suggesting the need for congestion control, scheduling and traffic management schemes. Second, source routing achieves higher throughputs while also generating significantly less routing overhead than AODV. Results also show that in some instances, the fairness of the network is very uneven among concurrent TCP flows, resulting in several sending stations achieving very little or no throughput.

High performance TCP in ANSNET

This report concentrates on specific requirements and goals of the research networks supported by ANSNET, but applies to any TCP dominated high speed WAN and in particular those striving to support high speed end-to-end flows. Measurements have been made under conditions intended to better understand performance barriers imposed by network equipment queueing capacities and queue drop strategies.The IBM RS/6000 based routers currently supporting ANSNET performed very well in these tests. Measurements have been made with the current software and performance enhanced software. Single TCP flows are able to achieve 40 Mb/s and competing multiple TCP flows achieve over 41 Mb/s link utilization on 44.7 Mb/s DS3 links with delays comparable to US cross continent ANSNET delays. Congestion collapse is demonstrated with intentionally reduced queueing capacity and using window sizes much larger than optimal.A variation of Floyd and Jacobson's Random Early Detection (RED) algorithm [1] is tested. Performance improved with the use of RED for tests involving multiple flows. With RED and queueing capacity at or above the delay bandwidth product, congestion collapse is avoided, allowing the maximum window size to safely be set arbitrarily high.Queueing capacity greater than or equal to the delay bandwidth product and RED are recommended. RED provides performance improvement in all but the single flow case, but cannot substitute for adequate queueing capacity, particularly if high speed flows are to be supported.