Dynamic Modeling of Internet tra¢c: Linear versus nonlinear canonical correlation analysis of HTTP versus FTP tra¢c

The Hypertext Transfer Protocol (HTTP) and the File Transfer Protocol (FTP) are both application-level protocols layered over the Transmission Control Protocol (TCP). HTTP is a request/response protocol used for the data transfer over the Internet. The primary function of FTP, on the other hand, is de...ned as transferring ...les e¢ciently and reliably among hosts. We study the statistical and dynamical properties of both HTTP and FTP tra¢c. We use Network Simulator (NS) to generate synthesized data for HTTP and FTP traf...c. We use the Canonical Correlation Analysis (CCA) as our statistical analysis tool. Both Linear Canonical Correlation Analysis (LCCA) and Nonlinear Canonical Correlation Analysis (NLCCA) are used to investigate deterministic, “chaotic,” and stochastic features of the signals. Comparison of linear predictor models and nonlinear predictor models suggests that the tra¢c signals possesses nonlinear features. Finally, NLCCA is implemented in a speci...c way, referred to as Alternating Conditional Expectation (ACE).

[1]  J. Baillieul,et al.  Identification and filtering of nonlinear systems using canonical variate analysis , 1990, 29th IEEE Conference on Decision and Control.

[2]  W. E. Larimore,et al.  Generalized canonical variate analysis of nonlinear systems , 1988, Proceedings of the 27th IEEE Conference on Decision and Control.

[3]  Stephen Eubank,et al.  State space forecasting and noise reduction , 1990, 29th IEEE Conference on Decision and Control.

[4]  Richard Coppola,et al.  Resonant neuroelectric patterns evoked by a cognitive task , 1990, 29th IEEE Conference on Decision and Control.

[5]  D. P. Martin.  Canonical variate analysis of large scale systems on a transputer array , 1990, 29th IEEE Conference on Decision and Control.

[6]  Jonathan B. Postel,et al.  RFC 959: File transfer protocol , 1985 .

[7]  Wallace E Larimore,et al.  System Identification and Filtering of Nonlinear Controlled Markov Processes by Canonical Variate Analysis , 1989 .

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

[9]  Duncan A. Mellichamp,et al.  Identification of chemical processes using canonical variate analysis , 1990, 29th IEEE Conference on Decision and Control.

[10]  P. Pruthi,et al.  Heavy-tailed on/off source behavior and self-similar traffic , 1995, Proceedings IEEE International Conference on Communications ICC '95.

[11]  J. Friedman,et al.  Estimating Optimal Transformations for Multiple Regression and Correlation. , 1985 .

[12]  Kihong Park.  On the relationship between le sizes, transport protocols, and self-similar network tra c , 1996 .

[13]  Sally Floyd,et al.  Wide-Area Traffic: The Failure of Poisson Modeling , 1994, SIGCOMM.

[14]  Alan S. Willsky,et al.  The Modeling and Estimation of Statistically Self-Similar Processes in a Multiresolution Framework , 1999, IEEE Trans. Inf. Theory.

[15]  Roy T. Fielding,et al.  Hypertext Transfer Protocol - HTTP/1.0 , 1996, RFC.

[16]  W. Linde.  STABLE NON‐GAUSSIAN RANDOM PROCESSES: STOCHASTIC MODELS WITH INFINITE VARIANCE , 1996 .

[17]  H. Hotelling.  Relations Between Two Sets of Variates , 1936 .

[18]  강인혜,et al.  [서평]High-Speed Networks : TCP/IP and ATM Design Principles , 1999 .

[19]  Wallace E. Larimore,et al.  Canonical variate analysis in identification, filtering, and adaptive control , 1990, 29th IEEE Conference on Decision and Control.

[20]  Wallace E. Larimore,et al.  Predictive inference, sufficiency, entropy and an asymptotic likelihood principle , 1983 .