Fairness and Stability Analysis of High Speed Transport Protocols
Balázs Sonkoly
Ph.D. Dissertation, 2009.
Abstract
TCP congestion control had managed successfully the stability of the Internet in the past decades but it has reached its limitations in "challenging" network environments. The new challenges of next-generation networks (e.g., high speed communication or the communication over different media) generated an urgent need to further develop the congestion control of the current Internet. In recent years, several new proposals and modifications of the standard congestion control mechanism have been developed by different research groups all over the world. These new mechanisms and TCP versions address different aspects of future networks and applications and improve the performance of regular TCP. One of the important network environments where the serious drawbacks of standard TCP Reno can be experienced is high speed wide are networks. These networks can be characterized by high bandwidth-delay product (BDP) and TCP cannot efficiently utilize them due to its conservative congestion control scheme. As a response to this problem, the research community has proposed several new transport protocols recently referred as high speed TCPs.
On the one hand, this dissertation provides a comprehensive fairness analysis of promising TCP proposals and congestion control mechanisms designed for high bandwidth-delay product networks. We put emphasis not only on the equilibrium behavior but also on the transient characteristics with the dynamic behavior to gain an in-depth insight into the interaction and co-existence of different congestion control schemes. We reveal the analytic properties of the transient and dynamic behavior of individual flows of different TCP protocols and define novel metrics in order to characterize them. The dynamic characteristics of competing flows are also investigated in a wide range of network scenarios. We investigate how the transient state can influence the long-term fairness performance. As part of this transient analysis, we deeply analyze the impacts of starting time on the performance and fairness. Here, the main focus is on inter-protocol properties when different TCP versions are interacting. The explanations of the experienced phenomena are also given based on the novel metrics and methodology.
On the other hand, a comprehensive control-theoretic analysis of HSTCP is also carried out to estimate the performance of the protocol in a very high bandwidth-delay product network environment where the queues are regulated by the RED active queue management. The motivation behind our approach is to gain analytical insight into the performance of HSTCP. To achieve this goal, firstly, we establish a fluid-flow model for HSTCP/RED networks. Secondly, we describe a systematic implementation methodology in detail in order to make the non-linear model tractable. The results of the model are validated by packet-level simulations. Thirdly, a stability condition for HSTCP/RED networks is derived which can be applied in RED design and parameter setting.
[Dissertation][Summary of the Dissertation]
A Comprehensive TCP Fairness Analysis in High Speed Networks
S. Molnár, B. Sonkoly, T. A. Trinh
Computer Communications, Elsevier, Volume 32, Issues 13-14, 17 August 2009, pp. 1460-1484.
Abstract
The short-term dynamics of competing high speed TCP flows can have strong impacts on their long-term fairness. This leads to severe problems for both the co-existence and the deployment feasibility of different proposals for the next generation networks. However, to our best knowledge, no root-cause analysis of this observation is available. This is the major motivation of our work.
The contribution of the paper is twofold. First, we present our comprehensive performance evaluation results of both inter- and intra-protocol fairness behavior of different TCP versions to get an overall view of these protocols. The analysis has revealed not only the equilibrium behavior but also the transient characteristics besides the dynamic behavior. Second, we show the results of a root-cause analysis to get a deeper understanding in the case of some promising TCP versions. This study does not only fill the "black holes", i.e. answers the questions which remained unanswered in some cases, but rather goes deeper and investigates questions which have never been asked yet. The work includes flow-level, packet-level, queueing and also spectral analyses. Three loss-based (HighSpeed TCP, Scalable TCP and BIC TCP) proposals and the delay-based FAST TCP are investigated in details with both "dumb-bell" and "parking-lot" topologies.
[link]Understanding the Impacts of Short-Term Dynamics on Long-Term Fairness of Competing High Speed TCP Flows: A Root-Cause Analysis
B. Sonkoly, T. A. Trinh, S. Molnár
NETWORKS 2008, Budapest, Hungary, September 28- October 2, 2008.
Abstract
The short-term dynamics of competing high speed TCP flows can have strong impacts on their long-term fairness. This leads to severe problems for both the co-existence and the deployment feasibility of different proposals for the next generation networks. However, to our best knowledge, no root-cause analysis of the observation is available. In this paper, we try to fill this gap by providing an in-depth root-cause analysis of this phenomenon. We demonstrate that the widely used Jain's index as a fairness metric can not provide sufficient characterization of the phenomena. More precisely, Jain's index does not reflect the dynamic flow behaviors, e.g., starting time of the flows. We provide an analytical and simulation study to show the importance of the flow dynamics on fairness. We also propose a new metric called saturation time for fairness characterization. Both AIMD-based (HighSpeed TCP, BIC TCP) and MIMD-based (Scalable TCP) TCP versions are investigated in different topologies, namely dumb-bell and parking-lot topologies. In extreme cases, we also analyze and explain the ``starving'' effect of competing high speed TCP flows, when a flow forces other flows to deviate from their proper operation.
[pdf]Revisiting FAST TCP Fairness
T. A. Trinh, B. Sonkoly, S. Molnár
18th ITC Specialist Seminar on Quality of Experience, Karlskrona, Sweden, May 29-30, 2008.
Abstract
Fairness of competing TCP flows is an integral and indispensable part of transport protocol design for next generation high bandwidth-delay product networks. In this paper we revisit FAST TCP fairness behavior based on a comprehensive performance evaluation study. We demonstrate that FAST TCP with proper parameter settings can always achieve fair behavior with HighSpeed TCP and Scalable TCP. We also show that this behavior is rather robust property of the protocol concerning different traffic mix or network topology. The dynamic behavior of reaching the fair equilibrium state can be different which is demonstrated in the paper. Our study also emphasizes the important need for finding a dynamic sensitive fairness metric for performance evaluation of transport protocols for next generation high bandwidth-delay product networks.
[pdf]A Comprehensive TCP Fairness Analysis in High Speed Networks
S. Molnár, B. Sonkoly and T. A. Trinh
Technical Report, Budapest University of Technology and Economics, 2007.
Abstract
The short-term dynamics of competing high speed TCP flows could have surprising impacts on their long-term fairness. As a result, this could have a severe impact on the co-existence and, finally, the deployment feasibility of different seemingly promising proposals for the next generation networks. However, to our best knowledge, no root-cause analysis of the observation is available. This is the major motivation of our work.
The contribution of the paper is twofold. First, we present our comprehensive performance evaluation results of both inter- and intra-protocol fairness behavior of different TCP versions to get an overall view of these protocols. The analysis has revealed not only the equilibrium behavior but also the transient characteristics with the dynamic behavior. Second, we have performed a root-cause analysis to get a deeper understanding in the case of some of the promising TCP versions. This study not only fills the "black holes", the questions which remained unanswered in some cases but rather goes deeper and investigates questions which have never been asked yet. The analysis spans multiple dimensions: flow-level, packet-level, queueing and spectral analysis. Three loss-based (HighSpeed TCP, Scalable TCP and BIC TCP) approaches and the delay-based FAST are investigated in details with both dumb-bell and parking-lot topologies.
[pdf]Modelling the Inter-operation of High Speed TCP Protocols
B. Simon, B. Sonkoly, S. Molnár
TRANSCOM 2007, Žilina, Slovak Republic, 25-27 June 2007.
Abstract
In this paper, we give the fluid-flow models of the network elements such as the TCP sources, the bottleneck queue and the RED module. It is important to emphasize that we used RED, because with Drop Tail we would have got different results. A MATLAB/Simulink environment is also designed and implemented to solve the analytically not tractable differential equations invoking numerical approximations. In this environment, the interaction of different flows is examined. The results of the models are validated by Ns-2 simulations. As a result, a more exact and detailed knowledge about the essential qualities, advantages and disadvantages of the high speed protocols are gained from the analysis.
Understanding HighSpeed TCP: A Control-Theoretic Perspective
B. Sonkoly, T. A. Trinh and S. Molnár
The Third IASTED International Conference on Communications and Computer Networks (CCN 2005), Marina del Rey, CA, USA, October 24-26, 2005.
Abstract
One of the most promising solution for transport protocol over very high bandwidth-delay product networks is HighSpeed TCP. However, little is known about its performance as well as its interaction with other elements of the network (such as the RED queue management). In this paper, a comprehensive control-theoretic analysis of HighSpeed TCP is provided. We develop a fluid-flow model of the HighSpeed TCP/RED network and use it to study its performance. The main contributions of this paper are the following. Firstly, we provide a fluid-flow model for HighSpeed TCP/RED networks. Secondly, a comprehensive and systematic implementation methodology is described in detail. We also provide a Simulink-based framework for analyzing fluid-based models. Thirdly, we give stability conditions for HighSpeed TCP/RED networks. Finally, the results are validated by simulations using Ns-2.
[pdf]Other Papers
http://hsnlab.tmit.bme.hu/~molnar