AI帮你理解科学

AI 生成解读视频

AI抽取解析论文重点内容自动生成视频


pub
生成解读视频

AI 溯源

AI解析本论文相关学术脉络


Master Reading Tree
生成 溯源树

AI 精读

AI抽取本论文的概要总结


微博一下
In this paper we present a controltheoretic approach to determine how a conversation can satisfy its throughput and queueing delay requirements by adapting its data transfer rate to changes in network state, and to prove that such adaptations do not lead to instability

A control-theoretic approach to flow control.

Computer Communication Review, no. 1 (1995): 188-201

被引用1088|浏览35
EI
下载 PDF 全文
引用
微博一下

摘要

This paper presents a control-theoretic approach to reactive flow control in networks that do not reserve bandwidth. We assume a round-robin-like queue service discipline in the output queues of the network’s switches, and propose deterministic and stochastic models for a single conversation in a network of such switches. These models mot...更多

代码

数据

0
简介
  • As networks move towards integrated service, there is a need for network control mechanisms that can provide users with different qualities of service in terms of throughput, delay and delay jitter [1].
  • A control theoretic approach to flow control requires that changes in the network state be observable.
  • A source sending data to a RAS server should use a rate-based flow control scheme that determines the allocated service rate, and sends data at this rate.
重点内容
  • As networks move towards integrated service, there is a need for network control mechanisms that can provide users with different qualities of service in terms of throughput, delay and delay jitter [1]
  • In this paper we present a controltheoretic approach to determine how a conversation can satisfy its throughput and queueing delay requirements by adapting its data transfer rate to changes in network state, and to prove that such adaptations do not lead to instability
  • A control theoretic approach to flow control requires that changes in the network state be observable
  • We have shown that it is possible to measure network state if the servers at the output queues of the switches are of a type called a Rate Allocating Server and the transport protocol uses the Packet-Pair probing technique [6,7,8]
  • The design strategy for the flow control mechanism is based upon the Separation Theorem [17]
  • In section 10.2, we show how windowbased flow control can be used in conjunction with a rate-based approach
结果
  • The authors explicitly model the dynamics of μ and so the control scheme can depend on the currently measured value of μ, as opposed to only on an asymptotic time average.
  • In the analysis, the authors treat the arrival of the packet pair as a single event, that measures both the round trip time and the bottleneck service rate.
  • The constant for this averager, β, is obtained from another fuzzy controller that links the change in error to the value of β.
  • Note that nb (k + 1), the estimate for the number of packets in the bottleneck queue, plays a critical role in the control system.
  • Since this packet has no queueing delays, the authors can estimate the propagation delay of the conversation from this packet’s measured round trip time.
  • With information about the propagation delay R, control can be done as quickly as once every packet-pair with no change to the length of the state vector.
  • This section considers two practical considerations: how to correct for parameter drift; and how to coordinate rate-based and window-based flow control.
  • A conversation is allocated a flow control window that is always larger than the product of the allocated bandwidth at the bottleneck, and the round trip propagation delay.
  • One body of work has considered the dynamics of a system where users update their sending rate either synchronously or asynchronously in response to measured round trip delays, or explicit congestion signals, for example in references [34,35,36,37,38].
结论
  • Their modeling of the service rate μ is as a random variable as opposed to a random walk, and though they propose the use of recursive minimum mean squared error filters to estimate system state, the bulk of the results assume complete information about the network state.
  • The authors have described two provably stable rate-based flow control schemes as well as a novel estimation scheme using fuzzy logic.
  • The performance of the flow control with Fair Queueing servers in the benchmark suite described in reference [10] is comparable to that of the DECbit scheme [47], but without any need for switches to set bits.
总结
  • As networks move towards integrated service, there is a need for network control mechanisms that can provide users with different qualities of service in terms of throughput, delay and delay jitter [1].
  • A control theoretic approach to flow control requires that changes in the network state be observable.
  • A source sending data to a RAS server should use a rate-based flow control scheme that determines the allocated service rate, and sends data at this rate.
  • The authors explicitly model the dynamics of μ and so the control scheme can depend on the currently measured value of μ, as opposed to only on an asymptotic time average.
  • In the analysis, the authors treat the arrival of the packet pair as a single event, that measures both the round trip time and the bottleneck service rate.
  • The constant for this averager, β, is obtained from another fuzzy controller that links the change in error to the value of β.
  • Note that nb (k + 1), the estimate for the number of packets in the bottleneck queue, plays a critical role in the control system.
  • Since this packet has no queueing delays, the authors can estimate the propagation delay of the conversation from this packet’s measured round trip time.
  • With information about the propagation delay R, control can be done as quickly as once every packet-pair with no change to the length of the state vector.
  • This section considers two practical considerations: how to correct for parameter drift; and how to coordinate rate-based and window-based flow control.
  • A conversation is allocated a flow control window that is always larger than the product of the allocated bandwidth at the bottleneck, and the round trip propagation delay.
  • One body of work has considered the dynamics of a system where users update their sending rate either synchronously or asynchronously in response to measured round trip delays, or explicit congestion signals, for example in references [34,35,36,37,38].
  • Their modeling of the service rate μ is as a random variable as opposed to a random walk, and though they propose the use of recursive minimum mean squared error filters to estimate system state, the bulk of the results assume complete information about the network state.
  • The authors have described two provably stable rate-based flow control schemes as well as a novel estimation scheme using fuzzy logic.
  • The performance of the flow control with Fair Queueing servers in the benchmark suite described in reference [10] is comparable to that of the DECbit scheme [47], but without any need for switches to set bits.
相关工作
  • Related Work and Contributions

    Several control theoretic approaches to flow control have been studied in the past. One body of work has considered the dynamics of a system where users update their sending rate either synchronously or asynchronously in response to measured round trip delays, or explicit congestion signals, for example in references [34,35,36,37,38]. These approaches typically assume Poisson sources, availability of global information, a simple flow update rule, and exponential servers. We do not make such assumptions. Further, they deal with the dynamics of the entire system, with the sending rate of all the users explicitly taken into account. In contrast, we consider a system with a single user, where the effects of the other users are considered as a system ‘noise’. Also, in our approach, each user uses a rather complex flow update rule, based in part on fuzzy prediction, and so the analysis is not amenable to the simplistic approach of these authors.
基金
  • This research was supported by the National Science Foundation and the Defense Advanced Research Projects Agency (DARPA) under Cooperative Agreement NCR-8919038 with the Corporation for National Research Initiatives, by AT&T Bell Laboratories, Hitachi, Ltd., the University of California under a MICRO grant, and the International Computer Science Institute
引用论文
  • D. Ferrari, Client Requirements for Real-Time Communications Services, IEEE Communications Magazine 28, 11 (November 1990).
    Google ScholarLocate open access versionFindings
  • C. R. Kalmanek, H. Kanakia and S. Keshav, Rate Controlled Servers for Very High Speed Networks, Proc. Globecom 1990, December 1990, 300.3.1-300.3.9.
    Google ScholarLocate open access versionFindings
  • D. Ferrari and D. Verma, A Scheme for Real-Time Channel Establishment in Wide-Area Networks, IEEE J. on Selected Areas in Communications, April 1990.
    Google ScholarLocate open access versionFindings
  • S. J. Golestani, A Stop-and-Go Queueing Framework for Congestion Management, Proc. ACM SigComm 1990, September 1990, 8-18.
    Google ScholarLocate open access versionFindings
  • L. Zhang, A New Architecture for Packet Switching Network Protocols, PhD thesis, Massachusetts Institute of Technology, July 1989.
    Google ScholarFindings
  • S. Keshav, A. K. Agrawala and S. Singh, Design and Analysis of a Flow Control Algorithm for a Network of Rate Allocating Servers, in Protocols for High Speed Networks II, Elsevier Science Publishers/North-Holland, April 1991.
    Google ScholarFindings
  • S. Singh, A. K. Agrawala and S. Keshav, Deterministic Analysis of Flow and Congestion Control Policies in Virtual Circuits, Tech. Rpt.-2490, University of Maryland, June 1990.
    Google ScholarFindings
  • S. Keshav, The Packet Pair Flow Control Protocol, Tech. Rpt. 91-028, International Comp. Sci. Institute, Berkeley, CA 94704, May 1991.
    Google ScholarLocate open access versionFindings
  • A. Greenberg and N. Madras, How Fair is Fair Queueing?, Proc. Performance 90, 1990.
    Google ScholarLocate open access versionFindings
  • A. Demers, S. Keshav and S. Shenker, Analysis and Simulation of a Fair Queueing Algorithm, Journal of Internetworking Research and Experience, September 1990, 3-26;. also Proc. ACM SigComm, Sept. 1989, pp 112..
    Google ScholarLocate open access versionFindings
  • H. Zhang and S. Keshav, Comparison of Rate-Based Service Disciplines, Proc. ACM SigComm 1991, September 1991. also International Comp. Sci. Institute Tech. Rpt. 91-024, Berkeley, CA..
    Google ScholarLocate open access versionFindings
  • S. Tripathi and A. Duda, Time-dependent Analysis of Queueing Systems, INFOR 24, 3 (1978), 334-346.
    Google ScholarFindings
  • J. G. Waclawsky, Window Dynamics, PhD Thesis, University of Maryland, College Park, May 1990.
    Google ScholarFindings
  • J. G. Waclawsky and A. K. Agrawala, Dynamic Behavior of Data Flow within Virtual Circuits, Comp. Sci.-Tech. Rpt.-2250, University of Maryland, May 1989.
    Google ScholarFindings
  • B. D. O. Anderson and J. B. Moore, Optimal Filtering, Prentice Hall, 1979.
    Google ScholarFindings
  • K. K. Sabnani and A. N. Netravali, A High Speed Transport Protocol for Datagram/Virtual Circuit Networks, Proc. ACM SigComm 1989, September 1989, 146-157.
    Google ScholarLocate open access versionFindings
  • B. D. O. Anderson and J. B. Moore, Linear Quadratic Methods, Prentice Hall, 1990.
    Google ScholarFindings
  • D. Mitra and J. B. Seery, Dynamic Adaptive Windows for High Speed Data Networks: Theory and Simulations, Proc. ACM SigComm 1990, September 1990, 30-40.
    Google ScholarLocate open access versionFindings
  • D. Mitra, Asymptotically Optimal Design of Congestion Control for High Speed Data Networks, To Appear in IEEE Trans. on Communications, 1991.
    Google ScholarLocate open access versionFindings
  • S. Keshav, Congestion Control in Computer Networks, PhD thesis, University of California, Berkeley, August 1991.
    Google ScholarFindings
  • A. E. Ekberg, D. T. Luan and D. M. Lucantoni, Bandwidth Management: A Congestion Control Strategy for Broadband Packet Networks: Characterizing the ThroughputBurstiness Filter, Proc. ITC Specialist Seminar, Adelaide, 1989, paper no. 4.4.
    Google ScholarFindings
  • K. Ogata, Discrete Time Control Systems, Prentice Hall, 1987.
    Google ScholarFindings
  • G. C. Goodwin and K. S. Sin, Adaptive Filtering Prediction and Control, Prentice Hall, 1984.
    Google ScholarFindings
  • H. J. Zimmerman, in Fuzzy Set Theory and its Applications, Kluwer Academic Publishers, 1985.
    Google ScholarFindings
  • L. A. Zadeh, Outline of a New Approach to the Analysis of Complex Systems and Decision Processes, IEEE Trans. on Systems, Man and Cybernetics, 1973, 28-44.
    Google ScholarLocate open access versionFindings
  • G. Langari, Analysis and Design of Fuzzy Control Systems, PhD thesis (in preparation), University of California, Berkeley, 1991.
    Google ScholarFindings
  • L. A. Zadeh, Fuzzy Sets, Journal of Information and Control 8 (1965), 338-353.
    Google ScholarLocate open access versionFindings
  • P. S. Khedkar and S. Keshav, Fuzzy Prediction of Timeseries, Proc. IEEE Conference on Fuzzy Systems-92, March 1992.
    Google ScholarLocate open access versionFindings
  • C. Agnew, Dynamic Modeling and Control of Congestionprone Systems, Operations Research 24, 3 (1976), 400419.
    Google ScholarLocate open access versionFindings
  • D. Tipper and M. K. Sundareshan, Numerical Methods for Modeling Computer Networks under Nonstationary Conditions, JSAC 8, 9 (December 1990).
    Google ScholarLocate open access versionFindings
  • J. Bolot, Dynamical Behavior of Rate-Based Flow Control Mechanisms, Comp. Sci.-Tech. Rpt. 2279.1, University of Maryland, October 1989.
    Google ScholarLocate open access versionFindings
  • R. Jain, Myths About Congestion Management in HighSpeed Networks, Technical Report-726, Digital Equipment Corporation, October 1990.
    Google ScholarFindings
  • E. L. Hahne, C. R. Kalmanek and S. P. Morgan, Fairness and Congestion Control on a Large ATM Data Network with Dynamically Adjustable Windows, 13th International Teletraffic Congress, Copenhagen, June 1991.
    Google ScholarFindings
  • K. Bharath-Kumar and J. M. Jaffe, A New Approach to Performance-Oriented Flow Control, IEEE Trans. on Communication COM-29, 4 (April 1981), 427-435.
    Google ScholarLocate open access versionFindings
  • S. Shenker, A Theoretical Analysis of Feedback Flow Control, Proc. ACM SigComm 1990, September 1990, 156-165.
    Google ScholarLocate open access versionFindings
  • C. Douligeris and R. Majumdar, User Optimal Flow Control in an Integrated Environment, Proc. of the Indo-US Workshop on Systems and Signals, January 1988. Bangalore, India.
    Google ScholarLocate open access versionFindings
  • A. D. Bovopoulos and A. A. Lazar, Asynchronous Algorithms for Optimal Flow Control of BCMP Networks, Tech. Rpt. WUCS-89-10, Washington University, St. Louis, MO, February 1989.
    Google ScholarFindings
  • A. D. Bovopoulos and A. A. Lazar, Decentralized Algorithms for Optimal Flow Control, Proc. 25th Allerton Conference on Communications Control and Computing, October 1987. University of Illinois, Urbana-Champaign.
    Google ScholarLocate open access versionFindings
  • K. K. Ramakrishnan and R. Jain, Congestion avoidance in Computer Networks with a Connectionless Network Layer - Part II - An Explicit Binary Feedback Scheme, Technical Report-508, Digital Equipment Corporation, April 1987.
    Google ScholarLocate open access versionFindings
  • V. Jacobson, Congestion Avoidance and Control, Proc. ACM SigComm 1988, August 1988, 314-329.
    Google ScholarLocate open access versionFindings
  • K. Ko, P. P. Mishra and S. K. Tripathi, Predictive Congestion Control in High-Speed Wide-Area Networks, in Protocols for High Speed Networks II, Elsevier Science Publishers/North-Holland, April 1991.
    Google ScholarFindings
  • J. Filipiak, Modelling and Control of Dynamic Flows in Communication Networks, Springer-Verlag, 1988.
    Google ScholarFindings
  • F. Vakil, M. Hsiao and A. A. Lazar, Flow Control in Integrated Local Area Networks, Vol. 7, 1987.
    Google ScholarLocate open access versionFindings
  • F. Vakil and A. A. Lazar, Flow Control Protocols for Integrated Networks with Partially Observed Traffic, IEEE Transactions on Automatic Control 32, 1 (1987), 2-14.
    Google ScholarLocate open access versionFindings
  • T. G. Robertazzi and A. A. Lazar, On the Modeling and Optimal Flow Control of the Jacksonian Network, Performance Evaluation 5 (1985), 29-43.
    Google ScholarFindings
  • M. Hsiao and A. A. Lazar, Optimal Flow Control of MultiClass Queueing Networks with Partial Information, IEEE Transactions on Automatic Control 35, 7 (July 1990), 855860.
    Google ScholarLocate open access versionFindings
  • K. K. Ramakrishnan and R. Jain, A Binary Feedback Scheme for Congestion Avoidance in Computer Networks, ACM ACM Trans. on Comp. Sys. 8, 2 (May 1990), 158181.
    Google ScholarLocate open access versionFindings
  • L. Zhang, S. Shenker and D. D. Clark, Observations on the Dynamics of a Congestion Control Algorithm: The Effects of Two-Way Traffic, Proc. ACM SigComm 1991, September 1991.
    Google ScholarLocate open access versionFindings
  • C. R. Kalmanek, Xunet 2: A Nationwide Testbed in HighSpeed Networking, Comp. Sci. Tech. Rpt., March 1991, AT&T Bell Labs, 600 Mountain Ave. Murray Hill, NJ 07974.
    Google ScholarLocate open access versionFindings
您的评分 :
0

 

标签
评论
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn
小科