Modeling Short-Lived TCP Connections with Open Multiclass Queuing Networks
In this paper we develop an open multiclass queuing network model to describe the behavior of short-lived TCP connections sharing a common IP network for the transfer of TCP segments. The queuing network model is paired with a simple model of the IP network, and the two models are solved through an iterative procedure. The combined model needs as inputs only the primitive network parameters, and produces estimates of the packet loss probability, the round trip time, the TCP connection throughput, and of the average TCP connection completion time (that is, of the average time necessary to transfer a file with given size over a TCP connection). The model presentation is centered on TCP-Tahoe, but the model of TCP-Reno is also available and results are presented. The analytical performance predictions are validated against detailed simulation experiments in a realistic networking scenario, proving that the proposed modeling approach is accurate.
KeywordsLoss Probability Packet Loss Probability Congestion Avoidance Fast Retransmit Slow Start Phase
- 2.J. Padhye, V. Firoiu, D. Towsley, and J. Kurose, “Modeling TCP Throughput: A Simple Model and its Empirical Validation,” Proc. ACM SIGCOMM’98, Sept., 1998.Google Scholar
- 3.N. Cardwell, S. Savage, T. Anderson, “Modeling TCP Latency,” Infocom 2000, Tel Aviv, Israel, March 2000.Google Scholar
- 4.V. Mishra, W. B. Gong, D. Towsley, “Fluid-based Analysis of a Network of AQM Routers Supporting TCP Flows with and application to RED”, in Proc. SIGCOMM’ 2000, Aug. 28-Sept. 1 2000, Stockholm, Sweden.Google Scholar
- 5.C.V. Hollot, V Mishra, W. B. Gong, D. Towsley, “A Control Theoretic Analysis of RED,” in Proc. IEEE Infocom 2001, Anchorage, Alaska, USA, April 22–26, 2001.Google Scholar
- 7.M. Garetto, R. Lo Cigno, M. Meo, M. Ajmone Marsan, “A Detailed and Accurate Closed Queueing Network Model of Many Interacting TCP Flows,” in Proc. IEEE Infocom 2001, Anchorage, Alaska, USA, April 22–26, 2001.Google Scholar
- 8.M. Garetto, R. Lo Cigno, M. Meo, M. Ajmone Marsan, “Queuing Network Models for the Performance Analysis of Multibottleneck IP Networks Loaded by TCP Short Lived Connections,” Politecnico di Torino Tech. Rep. DE/RLC/2001-5, Politecnico di Torino, June 2001. Available at http://www.tlc-networks.polito.it/locigno/papers/de-rlc-01-5.ps
- 9.E. Alessio, M. Garetto, R. Lo Cigno, M. Meo, M. Ajmone Marsan, “Analytical Estimation of the Completion Time of Mixed NewReno and Tahoe TCP Traffic over Single and Multiple Bottleneck Networks,” Proc. IEEE Globecom 2001, San Antonio, Tx, USA, Nov. 25–29, 2001.Google Scholar
- 10.M. Garetto, R. Lo Cigno, M. Meo, M. Ajmone Marsan, “On the Use of Queueing Network Models to Predict the Performance of TCP Connections,” Proc. 2001 Tyrrhenian International Workshop on Digital Communications, Taormina (CT), Italy, Sept. 17‐20, 2001.Google Scholar
- 11.ns-2, network simulator (ver.2). LBL, http://www-mash.cs.berkeley.edu/ns.
- 13.W. Stevens, “TCP Slow Start, Congestion Avoidance, Fast Retransmit, and Fast Recovery Algorithms,” RFC 2001, IETF, Jan. 1997Google Scholar
- 15.M. Garetto, R. Lo Cigno, M. Meo, M. Ajmone Marsan, “Modeling Short-Lived TCP Connections with Open Multiclass Queuing Networks-Extended Version,” Politecnico di Torino ech. Rep. DE/RLC/2001-4, Politecnico di Torino, June 2001. Available at http://www.tlc-networks.polito.it/locigno/papers/de-rlc-01-4.ps