Results 1  10
of
64
Experimental Evaluation of TCP Protocols for HighSpeed Networks
"... In this paper we present experimental results evaluating the performance of the ScalableTCP, HSTCP, BICTCP, FASTTCP and HTCP proposals in a series of benchmark tests. In summary, we find that both ScalableTCP and FASTTCP consistently exhibit substantial unfairness, even when competing flows ..."
Abstract

Cited by 69 (2 self)
 Add to MetaCart
In this paper we present experimental results evaluating the performance of the ScalableTCP, HSTCP, BICTCP, FASTTCP and HTCP proposals in a series of benchmark tests. In summary, we find that both ScalableTCP and FASTTCP consistently exhibit substantial unfairness, even when competing flows share identical network path characteristics. ScalableTCP, HSTCP, FASTTCP and BICTCP all exhibit much greater RTT unfairness than does standard TCP, to the extent that long RTT flows may be completely starved of bandwidth. ScalableTCP, HSTCP and BICTCP all exhibit slow convergence and sustained unfairness following changes in network conditions such as the startup of a new flow. FASTTCP exhibits complex convergence behaviour.
On Linear Copositive Lyapunov Functions and the Stability of Switched Positive Linear Systems
"... We consider the problem of common linear copositive function existence for positive switched linear systems. In particular, we present a necessary and sufficient condition for the existence of such a function for switched systems with two constituent linear timeinvariant (LTI) systems. A number of ..."
Abstract

Cited by 55 (4 self)
 Add to MetaCart
We consider the problem of common linear copositive function existence for positive switched linear systems. In particular, we present a necessary and sufficient condition for the existence of such a function for switched systems with two constituent linear timeinvariant (LTI) systems. A number of applications of this result are also given.
Flow Scheduling for Endhost Multihoming
 In Proceedings of IEEE INFOCOM
, 2006
"... Fueled by the competing DSL and Cable technologies, residential broadband access has seen a significant spread in availability to the point that many users have a choice from several ISPs. At the same time, 802.11 networks have spread rapidly in the residential area, and it is common for neighbors t ..."
Abstract

Cited by 46 (3 self)
 Add to MetaCart
(Show Context)
Fueled by the competing DSL and Cable technologies, residential broadband access has seen a significant spread in availability to the point that many users have a choice from several ISPs. At the same time, 802.11 networks have spread rapidly in the residential area, and it is common for neighbors to be able to access each other's wireless routers. Endusers can leverage this diversity to improve their Internet connectivity at no additional cost by pooling all available Internet connections, both their own and their neighbors' via wireless. In this paper we present our design and evaluation of flow scheduling algorithms in PERM, a framework for practical endhost multihoming. PERM scheduler employs automated online analysis of the endusers ' networking behaviors, and exploits the recognized patterns to achieve highperformance scheduling at flow level. We verify our models of enduser's network traffic with large residential TCP traces. Based on these models we propose algorithms for scalable preprobing and hybrid flow scheduling. Intensive experiments in our prototype testbed show that PERM scheduler reduces the latency by up to 50% for lightvolume flows, and reduces the mean transmission time of heavyvolume flows by nearly 28% and 62% compared with a single Cable or DSL connection respectively. The PERM scheduler also outperforms algorithms for enterprise multihoming by up to 15% and 27% in mean transmission time for light and heavyvolume flows respectively.
TCPillinois: A loss and delaybased congestion control algorithm for highspeed networks
"... Abstract — We introduce a new congestion control algorithm for high speed networks, called TCPIllinois. TCPIllinois uses packet loss information to determine whether the window size should be increased or decreased, and uses queueing delay information to determine the amount of increment or decrem ..."
Abstract

Cited by 42 (3 self)
 Add to MetaCart
(Show Context)
Abstract — We introduce a new congestion control algorithm for high speed networks, called TCPIllinois. TCPIllinois uses packet loss information to determine whether the window size should be increased or decreased, and uses queueing delay information to determine the amount of increment or decrement. TCPIllinois achieves high throughput, allocates the network resource fairly, and is incentive compatible with standard TCP. We also build a new stochastic matrix model, capturing standard TCP and TCPIllinois as special cases, and use this model to analyze their fairness properties for both synchronized and unsynchronized backoff behaviors. We finally perform simulations to demonstrate the performance of TCPIllinois.
Using the 802.11e EDCF to Achieve TCP Upload Fairness over WLAN Links
 In Proc. of 3rd Int’l Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt
, 2005
"... We investigate the use of the 802.11e MAC EDCF to address transport layer unfairness in WLANs. A simple solution is developed that uses the 801.11e AIF S and CWmin parameters to ensure fairness between competing TCP uploads. An analytic model of TCP transport over the modified channel is developed i ..."
Abstract

Cited by 28 (0 self)
 Add to MetaCart
(Show Context)
We investigate the use of the 802.11e MAC EDCF to address transport layer unfairness in WLANs. A simple solution is developed that uses the 801.11e AIF S and CWmin parameters to ensure fairness between competing TCP uploads. An analytic model of TCP transport over the modified channel is developed in order to study the fairness properties of the proposed scheme. In addition to fairness between competing TCP flows, consideration is extended to include characteristics of TCP flows such as RTT unfairness and responsiveness and we observe that TCP flows with a wireless bottleneck link exhibit quite different properties from flows with a wired bottleneck. I.
Mean Field Markov Models of Wireless Local Area Networks
, 2009
"... In 1998, Giuseppe Bianchi introduced a mean field Markov model of the fundamental medium access control protocol used in Wireless Local Area Networks (WLANs). Due to the model’s intuitive appeal and the accuracy of its predictions, since then there has been a vast body of material published that ext ..."
Abstract

Cited by 10 (0 self)
 Add to MetaCart
(Show Context)
In 1998, Giuseppe Bianchi introduced a mean field Markov model of the fundamental medium access control protocol used in Wireless Local Area Networks (WLANs). Due to the model’s intuitive appeal and the accuracy of its predictions, since then there has been a vast body of material published that extends and analyzes models of a similar character. As the majority of this development has taken place within the culture and nomenclature of the telecommunications community, the aim of the present article is to review this work in a way that makes it accessible to probabilists. In doing so, we hope to illustrate why this modeling approach has proved so popular, to explain what is known rigorously, and to draw attention to outstanding questions of a mathematical nature whose solution would be of interest to the telecommunications community. For nonsaturated WLANs, these questions include rigorous support for its fundamental decoupling approximation, determination of the properties of the selfconsistent equations and the identification of the queueing stability region. 1
Buffer Sizing for 802.11 Based Networks
"... We consider the sizing of network buffers in 802.11 based networks. Wireless networks face a number of fundamental issues that do not arise in wired networks. We demonstrate that the use of fixed size buffers in 802.11 networks inevitably leads to either undesirable channel underutilization or unne ..."
Abstract

Cited by 8 (0 self)
 Add to MetaCart
We consider the sizing of network buffers in 802.11 based networks. Wireless networks face a number of fundamental issues that do not arise in wired networks. We demonstrate that the use of fixed size buffers in 802.11 networks inevitably leads to either undesirable channel underutilization or unnecessary high delays. We present two novel dynamic buffer sizing algorithms that achieve high throughput while maintaining low delay across a wide range of network conditions. Experimental measurements demonstrate the utility of the proposed algorithms in a production WLAN and a lab testbed.
Stochastic equilibria of AIMD communication networks
 SIAM JOURNAL OF MATRIX ANALYSIS
, 2006
"... In this paper tools are developed to analyse a recently proposed random matrix model of communication networks that employ additiveincrease multiplicativedecrease (AIMD) congestion control algorithms. We investigate properties of the Markov process describing the evolution of the window sizes of ..."
Abstract

Cited by 7 (4 self)
 Add to MetaCart
In this paper tools are developed to analyse a recently proposed random matrix model of communication networks that employ additiveincrease multiplicativedecrease (AIMD) congestion control algorithms. We investigate properties of the Markov process describing the evolution of the window sizes of network users. Using paracontractivity properties of the matrices involved in the model, it is shown that the process has a unique invariant probability, and the support of this probability is characterized. Based on these results we obtain a weak law of large numbers for the average distribution of resources between the users of a network. This shows that under reasonable assumptions such networks have a welldefined stochastic equilibrium. ns2 simulation results are discussed to validate the obtained formulae. (The simulation program ns2, or network simulator, is an industry standard for the simulation of Internet dynamics.)
Positive observers and dynamic outputfeedback controllers for interval positive linear systems
 IEEE TRANS. CIRCUITS AND SYSTEMS (I)
, 2008
"... This paper is concerned with the design of observers and dynamic outputfeedback controllers for positive linear systems with interval uncertainties. The continuoustime case and the discretetime case are both treated in a unified linear matrix inequality (LMI) framework. Necessary and sufficient ..."
Abstract

Cited by 7 (3 self)
 Add to MetaCart
This paper is concerned with the design of observers and dynamic outputfeedback controllers for positive linear systems with interval uncertainties. The continuoustime case and the discretetime case are both treated in a unified linear matrix inequality (LMI) framework. Necessary and sufficient conditions for the existence of positive observers with general structure are established, and the desired observer matrices can be constructed easily through the solutions of LMIs. An optimization algorithm to the error dynamics is also given. Furthermore, the problem of positive stabilization by dynamic outputfeedback controllers is investigated. It is revealed that an unstable positive system cannot be positively stabilized by a certain dynamic outputfeedback controller without taking the positivity of the error signals into account. When the positivity of the error signals is considered, an LMIbased synthesis approach is provided to design the stabilizing controllers. Unlike other conditions which may require structural decomposition of positive matrices, all proposed conditions in this paper are expressed in terms of the system matrices, and can be verified easily by effective algorithms. Two illustrative examples are provided to show the effectiveness and applicability of the theoretical results.
Window Flow Control: Macroscopic Properties from Microscopic Factors
, 2008
"... This paper studies window flow control focusing on bridging the gap between microscopic factors such as burstiness in subRTT timescales, and observable macroscopic properties such as steady state bandwidth sharing and flow level stability. Using new models, we analytically capture notable effects ..."
Abstract

Cited by 6 (6 self)
 Add to MetaCart
(Show Context)
This paper studies window flow control focusing on bridging the gap between microscopic factors such as burstiness in subRTT timescales, and observable macroscopic properties such as steady state bandwidth sharing and flow level stability. Using new models, we analytically capture notable effects of microscopic behavior on macroscopic quantities. For lossbased protocols, we calculate the loss synchronization rate for different flows and use it to quantitatively explain the unfair bandwidth sharing between paced and unpaced TCP flows. For delaybased protocols, we show that the ratios of round trip delays are critical to the stability of the system. These results deepen the fundamental understanding of congestion control systems. Packet level simulations are used to verify our theoretical claims.