Analysis of a Feedback Fluid Model for Heterogeneous TCP Sources

In this paper we consider a bottleneck link and buffer used by one or two fluid sources that are subject to feedback. The feedback is such that the model captures essential aspects of the behavior of the Transmission Control Protocol as used in the Internet. During overflow, the buffer sends negative feedback signals to the sources to indicate that the sending rate should be reduced. Otherwise the buffer sends positive signals so as to increase the rate. In this context we find closed form expressions for the solution of the one-source case. The two-source case extends the single-source model considerably: we can control the behavior and parameters of each source individually. This enables us to study the impact of these parameters on the sharing of links and buffers. For the two-source case we solve the related two-point boundary value problem in the stationary case. We also establish a numerically efficient procedure to compute the coefficients of the solution of the differential equations. The numerical results of this model are presented in an accompanying paper.

[1]  Frank Kelly,et al.  Charging and rate control for elastic traffic , 1997, Eur. Trans. Telecommun..

[2]  W. R. Scheinhardt,et al.  Markov-modulated and feedback fluid queues , 1998 .

[3]  Eitan Altman,et al.  Analysis of two competing TCP/IP connections , 2002, Perform. Evaluation.

[4]  Ivo J. B. F. Adan,et al.  Analysis of a single-server queue interacting with a fluid reservoir , 1998, Queueing Syst. Theory Appl..

[5]  Frank Kelly,et al.  Rate control for communication networks: shadow prices, proportional fairness and stability , 1998, J. Oper. Res. Soc..

[6]  Werner Scheinhardt.  Analysis of feedback fluid queues , 2000 .

[7]  P. Lancaster,et al.  The theory of matrices : with applications , 1985 .

[8]  Patrick Brown,et al.  Resource sharing of TCP connections with different round trip times , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[9]  Vidyadhar G. Kulkarni,et al.  Fluid models for single buffer systems , 1998 .

[10]  Michel Mandjes,et al.  Simple models of network access, with applications to the design of joint rate and admission control , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[11]  Eitan Altman,et al.  Fairness analysis of TCP/IP , 2000, Proceedings of the 39th IEEE Conference on Decision and Control (Cat. No.00CH37187).

[12]  Thomas Bonald,et al.  Comparison of TCP Reno and TCP Vegas via Fluid Approximation , 1999 .

[13]  A.I. Elwalid,et al.  Statistical multiplexing with loss priorities in rate-based congestion control of high-speed networks , 1994, IEEE Trans. Commun..

[14]  Charles R. Johnson,et al.  Matrix analysis , 1985, Statistical Inference for Engineers and Data Scientists.

[15]  Matthew Mathis,et al.  The macroscopic behavior of the TCP congestion avoidance algorithm , 1997, CCRV.

[16]  Thomas E. Stern,et al.  Analysis of separable Markov-modulated rate models for information-handling systems , 1991, Advances in Applied Probability.

[17]  SemkeJeffrey,et al.  The macroscopic behavior of the TCP congestion avoidance algorithm , 1997 .

[18]  Laurent Massoulié,et al.  Bandwidth sharing: objectives and algorithms , 2002, TNET.

[19]  Raj Jain,et al.  Analysis of the Increase and Decrease Algorithms for Congestion Avoidance in Computer Networks , 1989, Comput. Networks.

[20]  Mischa Schwartz,et al.  Broadband integrated networks , 1996 .

[21]  Michel Mandjes,et al.  Performance analysis of heterogeneous interacting TCP sources , 2001 .

[22]  D. Mitra,et al.  Stochastic theory of a data-handling system with multiple sources , 1982, The Bell System Technical Journal.

[23]  Bruno Sericola,et al.  A fluid queue driven by a Markovian queue , 1999, Queueing Syst. Theory Appl..