Paxson, Vern
Floyd, Sally
Abstract
Network arrivals are often modeled as Poisson processes for analytic
simplicity, even though a number of traffic studies have shown that
packet interarrivals are not exponentially distributed. We evaluate
fifteen wide-area traces, investigating a number of wide-area TCP
arrival processes (session and connection arrivals, FTPDATA connection
arrivals within FTP sessions, and TELNET packet arrivals) to
determine the error introduced by modeling them using Poisson
processes. We find that user-initiated TCP session arrivals, such
as remote-login and file-transfer, are well-modeled as Poisson
processes with fixed hourly rates, but that other connection arrivals
are less persuasively Poisson; that modeling TELNET packet
interarrivals as exponential grievously underestimates the burstiness
of.
|
Keywords
poisson
traffic
pareto
burstiness
long range correlations
self similar
Notes
Statistical analysis of traffic captured at two internet gateways. A
structural modeling explanation of how this may arise. Distributions
matched to pareto, lognormal etc. Dsicussion of long range correlations.
- User initiated connections are relatively close to poisson
- Computer or timer driven connections are not
- Self similarity on the scale of tens of seconds or more
- No natural length for bursts - occur on all scales
- Burstiness is not limited by mulitplexing - unlike Poisson
|
Related Papers
Bibtex
@ARTICLE {paxson.floyd_failure95,
author = {Paxson, Vern and Floyd, Sally},
title = {Wide-area traffic: the failure of {Poisson} modeling},
journal = {IEEE/ACM Transactions on Networking},
number = {3},
volume = {3},
pages = {226 -- 224},
url = "ftp://ftp.ee.lbl.gov/papers/poisson.ps.Z",
year = {1995}
}