Internet Application Characteristics
Application	Arrivals	Request Sizes	Response Sizes
WWW	Pareto [7] Heavy-tailed, but not as much as transfers	?	Pareto for file sizes > 1K [6] Heavy-tailed
Telnet Session	Poisson [1]	?	Fairly Fixed?
Telnet Data	Pareto [1] Heavy-tailed (Think time)	Packets: Log-normal [1] Bytes: Log-extreme [2]	?
FTP Session	Poisson [1]	Negligible	Fairly Fixed?
FTP Data	Bursty [1] Network-dependent (mget operations)	?	Pareto[1] Very Heavy Upper Tail [1]
SMTP	Not Poisson [1] (Mailing lists)	?	?
NNTP	Not Poisson [1] (Flooding)	?	?
X11 Session	Poisson? [1]	Negligible	?
X11Connection	Not Poisson [1]	?	?

Miscellaneous Sources
Source	Distribution
"Self-similar" sources	ON-OFF times with heavy tails [5,6] Inter-renewal times with infinite variance
VBR video	Gamma/Pareto [3]

Characteristics of Aggregate Traffic
Characteristic	Distribution
Aggregate Traffic (Self-Similar)	Fractional Gaussian Noise[5,6] Fractional ARIMA [5,6]
Sources generating self-similar traffic	Pareto [5,6] ON-OFF sources with inter-renewal times with infinite variance. (ON and/or OFF times have heavy-tails)
Observed WWW Throughput	Log-normal [4]

Notes: Intensity of self-similarity increases as the aggregate traffic level increases.

References
[1] Vern Paxson and Sally Floyd, "Wide-Area Traffic: The Failure of Poisson Modeling"
[2] Vern Paxson, "Empirically-Derived Analytic Models of Wide-Area TCP Connections"
[3] Mark Garrett and Walter Willinger "Analysis, Modeling and Generation of Self-Similar VBR Video Traffic"
[4] Hari Balakrishnan, Srinivasan Seshan, Mark Stemm, and Randy Katz, "Analyzing Stability in Wide-Area Network Performance"
[5] Will Leland, Murad S. Taqqu, Walter Willinger, and Daniel Wilson, "On the Self-Similar Nature of Ethernet Traffic (Extended Version)"
[6] Mark Crovella and Azer Bestavros, "Self-Similarity in World Wide Web Traffic: Evidence and Possible Causes"
[7] Carlos R. Cunha, Azer Bestavros, and Mark E. Crovella, "Characteristics of WWW client-based traces"