No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Running TCP over IS-2000
Abstract
We analyze the effects of TCP algorithms on data transmission over IS-2000 links. We give details of our experience in emulating TCP over IS-2000 and discuss how bit errors and bandwidth oscillation can disturb TCP performance. We show that significant TCP throughput increase can be achieved by modifying the default TCP (and RLP) settings especially in the case of some "finite burst" modes of data transmission.
... Bandwidth Oscillation effects such as spurious retransmissions were identified elsewhere (e.g., [30]) as factors that degrade throughput. There are research studies [52], [54], which show that in some cases Bandwidth Oscillation can be the single most important factor in reducing throughput. For fixed TCP parameters the achievable throughput depends on the pattern of Inamura, et al. ...
... The TCP Timestamps option [2] provides this capability, allowing the TCP sender to sample the RTT from every segment that is acknowledged. Using timestamps in the mentioned scenario leads to a more conservative TCP retransmission timer and reduces the risk of triggering spurious timeouts [45], [52], [54], [60]. ...
... It is well known (and has been shown with experimental data) that RFC 1144 [40] TCP header compression does not perform well in the presence of packet losses [43], [52]. If a wireless link error is not recovered, it will cause TCP segment loss between the compressor and decompressor, and then RFC 1144 header compression does not allow TCP to take advantage of Fast Retransmit Fast Recovery mechanism. ...
... TCP is the dominant transport layer protocol in the IP protocol suite, which carries most of the Internet traffic such as Web browsing, bulk file transfer and Telnet. TCP was initially designed for wired networks, and hence performs poorly in the presence of delay spikes which are especially more frequent in today's wireless mobile networks than in traditional wired network [1], [2], [3], [4], [5]. A sudden increase of the instantaneous round trip time (RT T ) beyond the sender's retransmission timeout value (RT O) causes retransmission ambiguity [6], [7], resulting in Spurious Timeouts 1 (ST) and Spurious Fast Retransmissions 2 (SFR) which produce serious end-to-end TCP performance penalty [4], [7]. ...
... Since our proposed model considers the effect of ST, it can assist in determining an appropriate value of RT O min . 2) There is an increasing research interest to study the interaction between TCP and lower layer protocols in wireless environments [4], [15], [2]. The settings of lower layer protocols, such as handoff schemes in Mobile IP and retransmission schemes at the link layer, have a non-trivial impact on the frequency of TCP spurious timeouts. ...
... In Eqn. (2), Z N P can be obtained from [13] as given in Eqn. (16), E(Z LDP ) has already been developed in Sec. ...
TCP has been found to perform poorly in the presence of spurious timeouts (ST) caused by delay spikes which are especially more frequent in today's wireless mobile networks than in traditional wired network. Because STs are generally considered to represent a transient state in wired networks, previous research did not consider the effects of ST on the steady state performance of TCP. In this paper, we propose an analytical model for the sending rate and throughput of TCP Reno as a function of packet error rate and characteristics of spurious timeouts. The proposed model has been validated against simulation results and has been found to be more accurate than previous models in the presence of spurious timeouts.
... Using TCP on top of the link-layer error recovery can increase end-to-end performance. Therefore link-layer and end-to-end recovery can co-exist; this can lead to inefficient interaction between the two layers of ARQ (Automatic Repeat reQuest) protocols [4,12,11] In this paper, we describe optimize the link characteristics to optimize the performance of current commercial TCP implementations and elucidate the optimum values of the link control parameter in W-CDMA (IMT-2000)[9] and the TCP parameter. ...
... Optimizing Link Layer Retransmission Farid [12] reported the relationships between TCP and link layer characteristics as related to CDMA2000 1xRTT. The periodic resource sharing at the base station means that the terminals see bandwidth oscillation, which causes excess TCP time-out, due to the delay-jitter, and thus poor TCP performance. ...
... We can see that throughput saturates at the window size of 48-64 Kbytes. This is important because many OSs set the default receive window size to [8][9][10][11][12][13][14][15][16] Kbytes. ...
We address the interaction between TCP and RLC (Radio Link Control), the link layer retransmission protocol in W-CDMA, a 3rd generation cellular wireless network technology, in evaluating TCP performance in following points; l) Supress delay-jitter in link layer to avoid excess retransmissions. 2) Clarify the trade-off between jitter-suppression and link utilization to improve TCP throughput. 3) Optimize the link layer and TCP parameters. We show how to find the TCP and link ARQ parameters that yield optimum overall performance; simulations and emulation confirm their effectiveness. The coexistence of ARQ and TCP can lead to inefficient interaction; the delay-jitter on the link layer may trigger spurious TCP retransmission. The solution is to suppress jitter on the link layer. To manage the trade-off between this suppression and link utilization, we optimized with link parameter.
... The available bandwidth of a bottleneck link is reduced from 300 kbps down to 10 kbps. Such rapid bandwidth changes can occur due to on-demand allocation and release of a high speed radio channel [30] . The link RTT is increased by several times which causes a spurious TCP timeout. ...
... Another result of [18] was that in a scenario with congestion losses Eifel can suffer from genuine timeouts, but using advanced loss recovery algorithms such as Reno-SACK [38][8] and Limited Transmit [4] improves the situation. A study of cdma-2000 reports a possibility of " bandwidth oscillation " with certain configuration options from the standard IS-2000 Rel.A [30]. Due to switching of a high-speed radio channel between several users link RTT increases above the estimate of the TCP retransmission timer triggering spurious TCP timeouts. ...
... This is because Eifel restores the congestion control state and gets more data outstanding in the network. Higher RTT in such a case makes the timer less prone to spurious timeouts [30]. In summary, using a conservative RTO such as suggested in [40] 6 with timestamps provides a sufficient protection against excessive spurious timeouts in many cases. ...
Delays on Internet paths, especially including wireless links, can be highly variable. On the other hand, a current trend for modern TCPs is to deploy a fine-grain retransmission timer with a lower minimum timeout value than 1 s suggested by RFC2988. Spurious TCP timeouts cause unnecessary retransmissions and congestion control back-off. The Eifel algorithm detects spurious TCP timeouts and recovers by restoring the connection state saved before the timeout. This paper presents an enhanced version of the Eifel response to spurious timeouts and illustrates its performance benefits on paths with a high delay-bandwidth product. The refinements concern the following issues (1) an efficient operation in presence of packet losses (2) appropriate restoration of congestion control, and (3) adapting the retransmit timer to avoid further spurious timeouts. In our simulations the Eifel algorithm on paths with a high delay-bandwidth product can increase throughput by up to 250% and at the same decrease the load on the network by 3%. The proposed response also shows adequate performance on heavily congested paths.
... Bandwidth Oscillation effects such as spurious retransmissions were identified elsewhere (e.g., [30]) as factors that degrade throughput. There are research studies [52], [54], which show that in some cases Bandwidth Oscillation can be the single most important factor in reducing throughput. For fixed TCP parameters the achievable throughput depends on the pattern of This section provides further details on a few example 2.5G/3G technologies. ...
... The TCP Timestamps option [2] provides this capability, allowing the TCP sender to sample the RTT from every segment that is acknowledged. Using timestamps in the mentioned scenario leads to a more conservative TCP retransmission timer and reduces the risk of triggering spurious timeouts [45], [52], [54], [60]. ...
... It is well known (and has been shown with experimental data) that RFC 1144 [40] TCP header compression does not perform well in the presence of packet losses [43], [52]. If a wireless link error is not recovered, it will cause TCP segment loss between the compressor and decompressor, and then RFC 1144 header compression does not allow TCP to take advantage of Fast Retransmit Fast Recovery mechanism. ...
This document describes a profile for optimizing TCP to adapt so that it handles paths including second (2.5G) and third (3G) generation wireless networks. It describes the relevant characteristics of 2.5G and 3G networks, and specific features of example deployments of such networks. It then recommends TCP algorithm choices for nodes known to be starting or ending on such paths, and it also discusses open issues. The configuration options recommended in this document are commonly found in modern TCP stacks, and are widely available standards-track mechanisms that the community considers safe for use on the general Internet.
... This combined delay and rate variability leads to undesirable interactions with TCP (e.g., spurious timeouts, bufferbloat, etc.) and poor bandwidth utilization. Recent research efforts focus on understanding the impact of delay/rate variability on TCP performance [13] and improving congestion-control protocols in the face of large variations [10, 17, 20, 23]. Khafizov et al. [13] study the performance of TCP over IS- 2000 networks and find that bandwidth oscillation significantly degrades TCP throughput. ...
... Recent research efforts focus on understanding the impact of delay/rate variability on TCP performance [13] and improving congestion-control protocols in the face of large variations [10, 17, 20, 23]. Khafizov et al. [13] study the performance of TCP over IS- 2000 networks and find that bandwidth oscillation significantly degrades TCP throughput. Assuming that TCP cannot adapt quickly enough to the delay and bandwidth fluctuations of cellular links, others try to model the variations, either deterministically [10] or dynamically based on stochastic control theory [20] and statistical methods [23] . ...
With the advent of high-speed cellular access and the overwhelming popularity of smartphones, a large percent of today's Internet content is being delivered via cellular links. Due to the nature of long-range wireless signal propagation, the capacity of the last hop cellular link can vary by orders of magnitude within a short period of time (e.g., a few seconds). Unfortunately, TCP does not perform well in such fast-changing environments, potentially leading to poor spectrum utilization and high end-to-end packet delay.
In this paper we revisit seminal work in cross-layer optimization in the context of 4G cellular networks. Specifically, we leverage the rich physical layer information exchanged between base stations (NodeB) and mobile phones (UE) to predict the capacity of the underlying cellular link, and propose nCQIC, a cross-layer congestion control design. Experiments on real cellular networks confirm that our capacity estimation method is both accurate and precise. A CQIC sender uses these capacity estimates to adjust its packet sending behavior. Our preliminary evaluation reveals that CQIC improves throughput over TCP by 1.08-2.89x for small and medium flows. For large flows, CQIC attains throughput comparable to TCP while reducing the average RTT by 2.38-2.65x.
... networks [5] ...
... In this section we present a brief review of the previous works and state the contribution of our work. In [5] and [6], the eects of TCP behavior on data transmission over IS-2000 wireless links are studied based on computer simulation. Based on the simulation results, it is shown that the NAK-based non-reliable ARQ recovery scheme is crucial to mitigate the incompatibilities of TCP/IP with lossy radio links and improve the performance of wireless TCP. ...
We provide the cross-layer analysis of wireless TCP systems. The effect of error correlation on the behavior of link retransmission strategy and the end-to-end throughput of TCP layer are investigated. Based on the cross-layer analysis, a refinement of link layer protocol is proposed by consciously utilizing the information of channel correlations, which leads to the performance improvement of wireless TCP systems.
... TCP interpreting all packet losses as being due to congestions results in the unnecessary triggering of the congestion slow start procedure. TCP performance in wireless environments has been shown nowhere near as efficient as in wired networks [2]. Furthermore, when multiple TCP connections are concentrated on a base station, it is likely that one connection will take more bandwidth than others due to random link errors. ...
TCP suffers from poor throughput performance in wireless networks. Furthermore, when multiple TCP connections compete at the base station, link errors and congestion lead to serious unfairness among the connections. Although the issue of TCP performance in wireless networks has attracted significant attention, most reports focus only on TCP throughput and assume that there is only a single connection in a congestion-free network. This paper studies the throughput and fairness of popular improvement mechanisms (the Snoop [8] and ELN [5]) and TCP variants with multiple TCP connections. Simulation results show that the improvement mechanisms under investigation are effective to improve TCP throughput in a wireless network. However, they cannot provide fairness among multiple TCP connections. From the studies presented, it is concluded that mechanisms to enhance TCP fairness are needed in wireless networks.
... The ON-OFF CBR connection provides a piecewise constant best-effort available bandwidth that is useful to test the algorithm reactivity under strenuous conditions. This scenario is particularly interesting since it reproduces the case of bandwidth oscillation that is present in 2.5G/3G wireless systems [18].Figure 9 shows that the sending rate of one Reno TCP source is bursty and oscillating.Figure 10 shows that TFRC exhibits large fluctuations of the sending rate even though the available bandwidth on the forward path is piecewise constant. Figure 11 reports analogous results obtained using the ARC algorithm. ...
The existing end-to-end TCP congestion control algorithm is well suited for applications that are not sensitive to delay jitter and abrupt changes of the transmission rate, such as FTP data transfer, but it is not recommended for delivering video data, whose perceived quality is sensitive to delay jitter and changes in the sending rate. In particular, the window-based control of Reno TCP congestion control causes burstiness in data transmission, which not only requires large buffers at the client side to provide a smooth playout but also may provoke bursts of lost packets difficult to recover via forward error correction techniques. This paper proposes an adaptive rate-based control (ARC) algorithm that striclty mimics the real-time dynamics of TCP and is based on an end-to-end mechanism to estimate the connection available bandwidth. Computer simulations using ns-2 have been developed to compare the ARC with the Reno TCP and with the TCP-Friendly Rate Control (TFRC) algorithm. Single- and multibottleneck scenarios in the presence of homogeneous and heterogeneous traffic sources have been considered. Simulations have shown that the ARC algorithm improves fairness and is friendly toward Reno. On the other hand, TFRC revealed itself not to be friendly toward Reno since it mimics only the long term behaviour of Reno TCP. Finally, simulations have shown that ARC remarkably improves the goodput with respect to TFRC and Reno in the presence of lossy links.
Categories and Subject Descriptors:
C.2.2 [Computer Communication Networks]: Network
Protocols
General Terms:
Algorithms, Design
Keywords:
Congestion control design - Rate-based control - RTP/UDP
... The effect of bandwidth oscillation in CDMA2000 on TCP was studied by Yavuz and Khafizov [167]. Due to switching of a high-speed radio channel between several users, the link RTT can increase above the estimate of the TCP retransmit timer with certain configuration options from the standard IS-2000 Rel.A [91]. Then, spurious TCP timeouts are triggered and the performance is reduced. ...
... This causes bandwidth oscillation and spurious retransmissions. In [23][53], the authors show that the bandwidth oscillation can be the most important factor in reducing throughput. To reduce spurious retransmission, Eifel algorithm was proposed in [20][21] and it recommends using time stamp option in TCP. ...
The objective of this research is to investigate and develop an efficient and seamless data transport protocol for a heterogeneous wireless mobile network. In next-generation network, most of heterogeneous wireless mobile networks will be combined and complementarily constitute a hierarchical network. To integrate different networks, many challenging issues should be solved. In this thesis, an efficient and seamless data transport mechanisms are explored. We investigate the problems that the current transport control protocol (TCP) will experience within the heterogeneous mobile network. In a heterogeneous network, a mobile host experiences drastic changes in network condition during a session. Traditional TCP struggles with abrupt network changes by intersystem handoff and cannot work efficiently in this environment. We propose a TCP scheme to be tailored to the heterogeneous mobile network to support seamless data transport. In the proposed scheme, a TCP is informed the impending handoff events and works differently based on a handoff type. Simulation results present the proposed algorithm improves throughput, stabilizes data transmission rapidly, and provides a seamless data transfer. We also propose an adaptive resource management scheme within a 3G cellular network based on a users priority level to reduce the call dropping and blocking rates. In a heterogeneous network, a network that provides smaller bandwidth may struggle with handed-off calls being served with a higher bandwidth. Therefore, a resource management algorithm should be defined so that an ongoing call is not dropped by a handoff and provides seamless data transfer. We propose an adaptive resource management scheme based on downgrading the quality of some existing services in a 3G cellular network. We analyze the system capacity, call blocking rate and call dropping rate of the proposed algorithm, and simulate the performance variation of the downgraded traffic. The results show that the proposed scheme increases system capacity, and decreases the call dropping rate at the cost of small delay of the downgraded data traffic. Ph.D. Committee Chair: John Copeland; Committee Member: Henry L Owen III; Committee Member: Gee-Kung Chang; Committee Member: Jun Xu; Committee Member: Krishna V Palem
... The standard TCP implementation is NewReno with SACK options enabled. Figure 4 shows the typical performance improvement using simulated 1xRTT in the presence of a bandwidth oscillation effect particular to the technology [14]. The TRL-PEP mitigates this behavior to clearly outperform other TCP implementations at different packet error rates. ...
TCP performance enhancement in wireless access networks is an important ongoing area of research. It is known that the hostile nature of the wireless channel and the mobile nature of wireless users interact adversely with standard TCP congestion control mechanisms [1], causing a drastic reduction in throughput. This article surveys a selection of different approaches to managing TCP performance over wireless links, and presents the results of simulation and field trial results of a novel TCP performance enhancing proxy over diverse cellular radio access technologies based on the GSM, cdma2000, and UMTS standards. The proposed TRL TCP performance enhancing proxy has the advantages of being completely transparent to both TCP endpoints and tunable to different access technologies, and frequently achieves the maximum throughput available from any of the studied radio access technologies.
Cloud computing is consistently proving to be the dominant architecture of the future, and mobile technology is the catalyst. By having the processing power and storage remotely accessible, the main focus of the terminal is now related to connectivity and user-interface. The success of cloud-based applications greatly depends on the throughput experienced by the end user, which is why transport protocols play a key role in mobile cloud computing. This chapter discusses the main issues encountered in cloud networks that affect connection-oriented transport protocols. These issues include, but are not limited to, large delay connections, bandwidth variations, power consumption, and high segment loss rates. To reduce these adverse effects, a set of proposed solutions are presented; furthermore, the advantages and disadvantages are discussed. Finally, suggestions are made for future mobile cloud computing transportlayer designs that address different aspects of the network, such as transparency, congestion-intensity estimation, and quality-of-service integration.
We are facing two major cause of performance degradation in TCP traffic over wireless network such as W-CDMA that utilizing link ARQ mechanisms; jitter and loss in streams of segments. For mitigating the degradations, we explored and achieved as follows; 1) Leveraging the link layer optimizations techniques and established standardization, we evaluated the contribution of Eifel for WCDMA network against jitter. As a result, we can recommend the use of STO avoidance mechanism, such as Eifel. Eifel improved the throughput by around 25 kbps at BLER 10% sample, even with unoptimized link parameters. This result will have substantially impact for network operation because if we could assume the presence of STO avoidance mechanism, the network can be optimized more toward link efficiency on the trade-off curve. 2) We proposed an improvement for the Fast Timeout algorithm with loss retransmission mechanism for strengthen in loss recovery. Simulation results yield the following; a sender using the proposed algorithms holds one half of the previous cwnd value and so avoids unnecessary throughput degradation even if duplicate ACKs arrive following a timeout and the retransmitted segment is lost. The proposed algorithm overcomes Fast timeout at most around 33 kbps of improvement in throughput for single set of recovery in our simulation settings.
We suggest a more conservative management of TCP's retransmit timer
and a more careful response of the TCP sender on duplicate ACKs. The
former reduces the risk of triggering spurious timeouts, while the
latter eliminates potentially unnecessary retransmits. Although, we
believe that these suggestions are permitted (although not
explicitly) by RFC2581 and RFC2988, they do not seem to be widely
known nor deployed in existing TCP implementations. We therefore want
to make implementers aware of these choices.
It is expected that the proportional fair (PF) scheduler will be used widely in cdma2000 1xEV-DO systems because it maximizes the sum of each user's utility, which is given by the logarithm of its average throughput. However, in terms of short-term average throughput, PF scheduler may lead to a large RTT variation. We analyze the impact of PF scheduler on TCP start-up behavior through NS-2 simulation. To show the impact of PF scheduling on TCP, we also analyze the packet transmission delay under the PF scheduling policy through mathematical model.
In order to support high data rate requirements and effectively manage the scarce wireless resources, additional bandwidth channels are allocated and taken away from mobile stations in 3G wireless data networks quite frequently. A TCP sender connected to the mobile, on seeing ACKs coming at a faster pace after additional bandwidth allocation, turns overtly optimistic and injects data into the network at a rate that might be excessive for an intermediate router, thereby leading to loss of multiple packets and subsequent prolonged recovery and periods of underutilization. In this work, we characterize this problem using an analytical model for losses based on continuous flow approximation as well as an extensive simulation setup. We also illustrate how bandwidth oscillations create more severe congestion than an increase in number of users to the extent that even RED algorithm is unable to check the sharp growth of queues. As a result, multiple packets are lost in a droptail fashion. We further demonstrate the dependence of congestion due to bandwidth allocation on the time during which mobiles' rates are increased and observe the degradation in performance for typical load scenarios. We also try to identify the boundary for stable operation of RED and finally present some possible methods for improving the performance.
With the continuous growth of mobile users and web-based wireless applications, the performance of accessing web services via wireless is becoming one of the key issues. Regular TCP has many problems to tackle wireless connections. Especially for 2.5G or 3G mobile users, the bandwidth is always limited to access various IP applications. In this paper, we describe the design, implementation and evaluation of a web accelerator. Our proposed web accelerator increases the web pages retrieval speed and maximizes bandwidth utilization of TCP/IP application to overcome the obstacles of wireless networks. This goal has been achieved by two solutions: first, replacing TCP with Boosted TCP Protocol (BTCP) to improve the performance of TCP over wireless; second, compressing the web content to make it efficient for various wireless devices . .
High Speed Downlink Packet Access (HSDPA) employed in 3G networks considerably augments wireless bandwidth, which enables a variety of mobile services based on TCP. However, the fundamental assumptions of TCP are disrupted by the properties hold by wireless network, such as high bit error rate, large delay and drastic bandwidth oscillation, which lead to the consequences of spurious time outs, spurious retransmission, and sending rate lagging behind the variation of bandwidth, degrading the wireless utilization greatly. To solve the problem, we design a cross-layer flow control algorithm which based on the bandwidth information of the current network environment and calculate a reference value to adjust the sending rate of the multiple senders. The experimental results show that the wireless bandwidth resource is sufficiently utilized, the queue length is stable and the sending rate is insensitive to the round trip time when using this algorithm.
To facilitate a viable evolution of cellular networks toward extensive packet data traffic, the High Speed Downlink Packet Access (HSDPA) technology is introduced. The various link adaptation techniques employed by HSDPA augment the bandwidth variation, which is identified as one of the most important factors resulting in the deterioration of TCP performance. In this paper, we firstly build an analytical model of TCP throughput to explain why the bandwidth variation degrades the TCP performance. Subsequently, a split-connection Window Adaptation TCP Proxy is proposed to improve the TCP throughput in HSDPA networks. To use the precious cellular link resources sufficiently, the length of the queue in Node-B is intentionally kept around the reference value through adaptively adjusting the sending window size of TCP proxy based on the dynamic values of varying bandwidth. Since both the disturbance caused by bandwidth variation and the feedback delay are prone to lead an unstable queue system, the robust sliding mode variable structure control theory is employed to design the proper control law to weaken the impact of noise and delay on the stability of the queue system in Node-B. The theoretical analysis and the enhanced scheme are verified through simulation experiments. The simulation results show that our TCP proxy is able to resist against bandwidth oscillation and improve the cellular link utilization.
Spurious TCP retransmission timeouts (RTOs) have been reported to be a problem on network paths involving links that are prone to sudden delays due to various reasons. Especially many wireless network technologies contain such links. Spurious retransmission timeouts often cause unnecessary retransmission of several segments, which is harmful for TCP performance. Recent proposals for avoiding unnecessary retransmissions after a spurious RTO require use of TCP options which must be implemented and enabled at both ends of teh connection. We introduce a new TCP sender algorithm for recovery after a retransmission timeout and show that unnecessary retransmissions after a spurious retransmission timeout, improving the TCP performance considerably. The algorithm is friendly towards other TCP connections, because it follows the congestion control principles and injects packets to the network at same rate as a conventional TCP sender. We implemented the algorithm and compared its performance to conventional TCP and Eifel TCP when RTOs occurred either due to sudden delays or due to packet losses. The results show that our algorithm either improves performance or gives similar througput as the other TCP variants evaluated in different test cases.
The objective of this research is to investigate and develop an efficient and seamless data transport protocol for a heterogeneous wireless mobile network. In next-generation network, most of heterogeneous wireless mobile networks will be combined and complementarily constitute a hierarchical network. To integrate different networks, many challenging issues should be solved. In this thesis, an efficient and seamless data transport mechanisms are explored. We investigate the problems that the current transport control protocol (TCP) will experience within the heterogeneous mobile network. In a heterogeneous network, a mobile host experiences drastic changes in network condition during a session. Traditional TCP struggles with abrupt network changes by intersystem handoff and cannot work efficiently in this environment. We propose a TCP scheme to be tailored to the heterogeneous mobile network to support seamless data transport. In the proposed scheme, a TCP is informed the impending handoff events and works differently based on a handoff type. Simulation results present the proposed algorithm improves throughput, stabilizes data transmission rapidly, and provides a seamless data transfer. We also propose an adaptive resource management scheme within a 3G cellular network based on a users priority level to reduce the call dropping and blocking rates. In a heterogeneous network, a network that provides smaller bandwidth may struggle with handed-off calls being served with a higher bandwidth. Therefore, a resource management algorithm should be defined so that an ongoing call is not dropped by a handoff and provides seamless data transfer. We propose an adaptive resource management scheme based on downgrading the quality of some existing services in a 3G cellular network. We analyze the system capacity, call blocking rate and call dropping rate of the proposed algorithm, and simulate the performance variation of the downgraded traffic. The results show that the proposed scheme increases system capacity, and decreases the call dropping rate at the cost of small delay of the downgraded data traffic. John Copeland, Committee Chair ; Henry L Owen III, Committee Member ; Gee-Kung Chang, Committee Member ; Krishna V Palem, Committee Member ; Jun Xu, Committee Member. Thesis (Ph. D.)--Electrical and Computer Engineering, Georgia Institute of Technology, 2006.
As a highly efficient radio link technology, the opportunistic scheduling policy has been widely investigated and implemented. Therefore, new challenges come up in the case of TCP over opportunistic scheduling systems. In this paper we investigate the impact of wireless opportunistic scheduling on TCP throughput. It shows that the optimization of the wireless link mechanisms needs to be maintained at the transport layer by cooperation of the adjacent layers. We propose a new ACK Reservoir Method to smooth the TCP behavior combating against the spurious timeout caused by scheduling. Based on performance analysis and simulation results, the proposed method is shown to enhance the TCP throughput by up to 100% in the presence of opportunistic scheduling. Finally, the proposed method is evaluated in a practical simulation scenario of CDMA/HDR system.
In order to support high data rate requirements and effectively manage scarce wireless resources, additional bandwidth channels are quite frequently allocated and taken away from mobile stations in 3G wireless data networks. A TCP sender connected to the mobile, on seeing ACKs coming at a faster pace after additional bandwidth allocation, turns overtly optimistic and injects data into the network in a more bursty manner that might be excessive for an intermediate router, thereby leading to loss of multiple packets and subsequent prolonged recovery and periods of underutilization. We characterize this problem using an analytical model for losses based on a continuous flow approximation as well as an extensive simulation setup. We also illustrate how bandwidth oscillations create more severe congestion than an increase in the number of users to the extent that even the RED algorithm is unable to check the sharp growth of queues. As a result, multiple packets are lost in a droptail fashion. We further demonstrate the dependence of congestion due to bandwidth allocation on the time during which mobiles' rates are increased and observe the degradation in performance for typical load scenarios.
The interpretation by transmission control protocol (TCP) of loss of linkage as loss due to congestion causes significant inefficiency in a wireless network. Congestion that often occurs in a wireless LAN results in further reduced performance. The paper introduces a method of controlling congestion in a network via a feedback mechanism called fair intelligent congestion control (FICC) which is useful for improving TCP performance in such a network. FICC overcomes the inefficiency of TCP in misinterpreting the link loss as congestion loss. Simulation testing has demonstrated that, by eliminating buffer overflows caused by congestion, FICC provides desirable fairness, reduces delay variation and improves effective throughput.
In a cdma2000 system, most user applications operate using protocols such as HTTP, FTP, and WAP that are run over TCP. Consequently, optimizing TCP performance over cdma2000 is essential for providing good user experience and for optimizing system capacity. In this paper, we study a number of issues that affect TCP performance, and show that TCP parameter optimization can, in a number of cases, have a significant impact on system performance. We finally suggest suitable TCP parameter values to use and other methods to optimize system throughput.
We introduce a new proxy that effectively prevents unnecessary retransmissions from flowing over a wireless link on a path with sudden delay. The proposed STD algorithm detects a spurious timeout based on the comparison of the measured inter-arrival timeout of ACK during the spurious timeout period with the average inter-arrival time of ACK. It responds to spurious timeout by filtering duplicate ACKs that can cause spurious fast retransmission. Simulation results show that the proposed spurious timeout detection (STD) algorithm performs better than, or as well as, other end-to-end mechanisms in the ranges of data rate which current wireless networks can provide.
We analyze the effects of TCP algorithms on data transmission over links with variable bandwidth. Significant throughput degradation is demonstrated due to factors such as spurious TCP retransmissions. The degradation is shown to be severe especially when high bandwidth resources are allocated to users for time periods during which multiple TCP segments are transmitted. The throughput degradation due to TCP effects is quantified for a wide range of burst and delay durations. A model for the TCP round trip time (RTT) is developed and validated with experimental results. This model is used to provide insight into spurious TCP retransmission problem and to suggest methods to alleviate it. Consequently, various methods are proposed and evaluated to improve the system throughput. Although we focus on IS-2000 links with finite burst mode of operation, the results would apply to different systems with shared resources and slow bandwidth allocation mechanisms.
Wireless data networks rely on link layer protocols to reduce relatively high frame error rates (FER) at the physical layer. One of the questions in wireless data networks design is how persistent a link-layer should be in performing retransmissions. This paper investigates this problem in the example of cdma2000 networks that employ the radio link protocol type 3 (RLP) for reducing physical layer FER. We develop an analytical model describing RLP transmission delay, residual FER and data throughput as a function of physical layer FER. We show how the model can be used to determine the dependency of TCP and RLP throughput on various RLP configuration parameters and to minimize RF cost. Model predictions are shown to agree with simulation results and were successfully applied to improving performance of a cdma2000 system.
The various link adaptation techniques employed by High Speed Downlink Packet Access (HSDPA) in the third generation (3G) networks augment the bandwidth oscillation, which is identified as one of the most important factors resulting in the throughput deterioration of Transmission Control Protocol (TCP). In this paper, we firstly explain why the bandwidth oscillation degrades the TCP performance through a special simulation experiment. Subsequently, a split connection Window Adaptation TCP Proxy is proposed to improve the TCP throughput over HSDPA networks. In this solution, the built-in attributes of HSDAP system are sufficiently utilized. In order to effectively use the precious cellular link resources, the length of the queue connected with it is intentionally kept around the reference value through adjusting the sending window size of TCP proxy based on the dynamic values of varying bandwidth. A discrete-time stochastic state space model is formulated to analyze the system stability. The validity of enhanced scheme is verified through simulation experiments. The performance of TCP proxy is compared with the standard TCP protocol. The numerical results show that our TCP proxy is able to keep the cellular link utilization over 90%, and to improve TCP throughput by 100% under most conditions.
In this letter, we introduce a new proxy that effectively prevents unnecessary retransmissions from flowing over a wireless link on a path with sudden delay. The proposed Spurious Timeout (STD) algorithm detects spurious timeout based on the data and acknowledge sequence number. It responses to spurious timeout by filtering duplicate acknowledgements that can cause spurious fast retransmission. Simulation results show that the proposed STD algorithm performs better than, or as well as, other end-to-end mechanisms.
We analyze two alternative retransmission timers for the Transmission Control Protocol (TCP). We first study the retransmission timer of TCP-Lite which is considered to be the current de facto standard for TCP implementations. After revealing four major problems of TCP-Lite's retransmission timer, we propose a new timer, named the Eifel retransmission timer, that eliminates these. The strength of our work lies in its hybrid analysis methodology. We develop models of both retransmission timers for the class of network-limited TCP bulk data transfers in steady state. Using those models, we predict the problems of TCP-Lite's retransmission timer and develop the Eifel retransmission timer. We then validate our model-based analysis through measurements in a real network that yield the same results.
this report, we illustrate some of the benefits of adding SACK to "Reno" implementations of TCP. We use simulations to show that a SACK option with at most three SACK blocks can be of substantial benefit, relative to TCP without SACK. We have been running simulations of Reno TCP with SACK in our simulator for over five years [Flo91]. Our assumption is that the important issue is that SACK be added to TCP, and that it is not of critical importance whether that addition is in the form of a SACK option with three SACK blocks, a SACK option with six SACK blocks, or other proposed formats (such as those suggested by Matt Mathis or Keshav). 2 TCP Background and Terminology
Abstract TCP,may,experience,poor,performance,when,multiple,packets,are,lost from,one window,of data.,With,the limited,information,available from cumulative acknowledgments,a TCP sender can only learn about a single,lost,packet,per round,trip,time. An aggressive,sender,could choose to retransmit packets early, but such retransmitted segments may,have,already,been,successfully,received. A Selective Acknowledgment (SACK) mechanism, combined with a
In October of '86, the Internet had the first of what became a series of 'congestion collapses'. During this period, the data throughput from LBL to UC Berkeley (sites separated by 400 yards and three IMP hops) dropped from 32 Kbps to 40 bps. Mike Karels ¹ and I were fascinated by this sudden factor-of-thousand drop in bandwidth and embarked on an investigation of why things had gotten so bad. We wondered, in particular, if the 4.3BSD (Berkeley UNIX) TCP was mis-behaving or if it could be tuned to work better under abysmal network conditions. The answer to both of these questions was “yes”.
Since that time, we have put seven new algorithms into the 4BSD TCP: round-trip-time variance estimation
exponential retransmit timer backoff
slow-start
more aggressive receiver ack policy
dynamic window sizing on congestion
Karn's clamped retransmit backoff
fast retransmit Our measurements and the reports of beta testers suggest that the final product is fairly good at dealing with congested conditions on the Internet.
This paper is a brief description of ( i ) - ( v ) and the rationale behind them. ( vi ) is an algorithm recently developed by Phil Karn of Bell Communications Research, described in [KP87]. ( viii ) is described in a soon-to-be-published RFC.
Algorithms ( i ) - ( v ) spring from one observation: The flow on a TCP connection (or ISO TP-4 or Xerox NS SPP connection) should obey a 'conservation of packets' principle. And, if this principle were obeyed, congestion collapse would become the exception rather than the rule. Thus congestion control involves finding places that violate conservation and fixing them.
By 'conservation of packets' I mean that for a connection 'in equilibrium', i.e., running stably with a full window of data in transit, the packet flow is what a physicist would call 'conservative': A new packet isn't put into the network until an old packet leaves. The physics of flow predicts that systems with this property should be robust in the face of congestion. Observation of the Internet suggests that it was not particularly robust. Why the discrepancy?
There are only three ways for packet conservation to fail: The connection doesn't get to equilibrium, or
A sender injects a new packet before an old packet has exited, or
The equilibrium can't be reached because of resource limits along the path. In the following sections, we treat each of these in turn.
Wireless data networks rely on link layer protocols to reduce relatively high frame error rates (FER) at the physical layer. One of the questions in wireless data networks design is how persistent a link-layer should be in performing retransmissions. This paper investigates this problem in the example of cdma2000 networks that employ the radio link protocol type 3 (RLP) for reducing physical layer FER. We develop an analytical model describing RLP transmission delay, residual FER and data throughput as a function of physical layer FER. We show how the model can be used to determine the dependency of TCP and RLP throughput on various RLP configuration parameters and to minimize RF cost. Model predictions are shown to agree with simulation results and were successfully applied to improving performance of a cdma2000 system.
Comparisons of Tahoe
- K Fall
- S Floyd
K. Fall, S. Floyd, " Comparisons of Tahoe, Reno and SACK TCP ", December 1995
work in progress: " TCP over 2.5G and 3G Wireless Networks " , draft-ietf-pilc-2
- H Inamura
H. Inamura, et. al., work in progress: " TCP over 2.5G and 3G Wireless Networks ", draft-ietf-pilc-2.5g3g-05
cdma2000 Standards for Spread Spectrum Systems
- Tia Eia
- Is
TIA/EIA/IS-2000, " cdma2000 Standards for Spread Spectrum Systems "
work in progress Advice for Internet Subnetwork Designers " , draft-ietf-pilc-link-arq-issues-08
- P Karn
P. Karn, et.al., work in progress: " Advice for Internet Subnetwork Designers ", draft-ietf-pilc-link-arq-issues-08