The origin of TCP traffic burstiness in short time scales

Internet traffic exhibits multifaceted burstiness and correlation structure over a wide span of time scales. Previous work analyzed this structure in terms of heavy-tailed session characteristics, as well as TCP timeouts and congestion avoidance, in relatively long time scales. We focus on shorter scales, typically less than 100-1000 milliseconds. Our objective is to identify the actual mechanisms that are mostly responsible for creating bursty traffic in those scales. We show that TCP selfclocking, joint with queueing in the network, can shape the packet interarrivals of a TCP connection in a two-level ON-OFF pattern. This structure creates strong correlations and burstiness in time scales that extend up to the Round-Trip Time (RTT) of the connection, especially for bulk transfers that have a large bandwidthdelay product relative to their window size. Also, the aggregation of many flows, without rescaling their packet interarrivals, does not converge to a Poisson stream, as one might expect from classical superposition results. Instead, the burstiness in those scales can be significantly reduced by TCP pacing, depending however on the minimum pacing timer. Finally, we show that sub-RTT burstiness is important in queueing performance not only in moderate load conditions, as previously shown, but also in high loads when the bottleneck buffer size is relatively small.

[1]  Michalis Faloutsos,et al.  A nonstationary Poisson view of Internet traffic , 2004, IEEE INFOCOM 2004.

[2]  Scott Shenker,et al.  Some observations on the dynamics of a congestion control algorithm , 1990, CCRV.

[3]  Donald F. Towsley,et al.  On the autocorrelation structure of TCP traffic , 2002, Comput. Networks.

[4]  Liang Guo,et al.  Corrections to "How Does TCP Generate Pseudo-Self-Similarity?" , 2002, CCRV.

[5]  Walter Willinger,et al.  Self-similarity through high-variability: statistical analysis of Ethernet LAN traffic at the source level , 1997, TNET.

[6]  Van Jacobson,et al.  Congestion avoidance and control , 1988, SIGCOMM '88.

[7]  Scott Shenker,et al.  Observations on the dynamics of a congestion control algorithm: the effects of two-way traffic , 1991, SIGCOMM '91.

[8]  Jie Wu,et al.  A unified framework for understanding network traffic using independent wavelet models , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[9]  Iraj Saniee,et al.  Performance impacts of multi-scaling in wide area TCP/IP traffic , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[10]  Richard G. Baraniuk,et al.  A Multifractal Wavelet Model with Application to Network Traffic , 1999, IEEE Trans. Inf. Theory.

[11]  Zhi-Li Zhang,et al.  Small-time scaling behaviors of Internet backbone traffic: an empirical study , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[12]  Walter Willinger,et al.  Self-Similar Network Traffic and Performance Evaluation , 2000 .

[13]  Jeffrey C. Mogul,et al.  Observing TCP dynamics in real networks , 1992, SIGCOMM '92.

[14]  Patrice Abry,et al.  Wavelet Analysis of Long-Range-Dependent Traffic , 1998, IEEE Trans. Inf. Theory.

[15]  Patrice Abry,et al.  Does fractal scaling at the IP level depend on TCP flow arrival processes? , 2002, IMW '02.

[16]  Richard G. Baraniuk,et al.  Connection-level analysis and modeling of network traffic , 2001, IMW '01.

[17]  Walter Willinger,et al.  On the self-similar nature of Ethernet traffic , 1993, SIGCOMM '93.

[18]  Patrice Abry,et al.  Cluster processes: a natural language for network traffic , 2003, IEEE Trans. Signal Process..

[19]  Ward Whitt,et al.  Characterizing Superposition Arrival Processes in Packet Multiplexers for Voice and Data , 1986, IEEE J. Sel. Areas Commun..

[20]  Hao Jiang,et al.  Source-level IP packet bursts: causes and effects , 2003, IMC '03.

[21]  Amit Aggarwal,et al.  Understanding the performance of TCP pacing , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[22]  Anja Feldmann,et al.  Data networks as cascades: investigating the multifractal nature of Internet WAN traffic , 1998, SIGCOMM '98.

[23]  Anja Feldmann,et al.  The changing nature of network traffic: scaling phenomena , 1998, CCRV.

[24]  Anja Feldmann,et al.  Dynamics of IP traffic: a study of the role of variability and the impact of control , 1999, SIGCOMM '99.

[25]  Donald F. Towsley,et al.  Modeling, simulation and measurements of queuing delay under long-tail internet traffic , 2003, SIGMETRICS '03.

[26]  Azer Bestavros,et al.  Self-similarity in World Wide Web traffic: evidence and possible causes , 1996, SIGMETRICS '96.

[27]  Hao Jiang,et al.  Passive estimation of TCP round-trip times , 2002, CCRV.

[28]  Hao Jiang,et al.  The Effect of Flow Capacities on the Burstiness of Aggregated Traffic , 2004, PAM.

[29]  Raj Jain,et al.  Packet Trains-Measurements and a New Model for Computer Network Traffic , 1986, IEEE J. Sel. Areas Commun..

[30]  Sándor Molnár,et al.  Queueing performance estimation for general multifractal traffic , 2003, Int. J. Commun. Syst..

[31]  Jin Cao,et al.  Internet Traffic Tends Toward Poisson and Independent as the Load Increases , 2003 .

[32]  Walter Willinger,et al.  Experimental queueing analysis with long-range dependent packet traffic , 1996, TNET.

[33]  Matthias Grossglauser,et al.  On the relevance of long-range dependence in network traffic , 1996, SIGCOMM '96.

[34]  Robert Tappan Morris,et al.  Variance of aggregated Web traffic , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).