Title :
Estimation of flow distributions tails from sampled traffic
Author :
Antunes, Nelson ; Pipiras, Vladas
Author_Institution :
Dept. of Math., Univ. of Algarve, Faro, Portugal
Abstract :
This paper addresses the problem of estimating the tails of flow duration and size distributions under various sampling methods. For each method the asymptotic equivalence between the flow duration distribution tail and the sampled flow quantities is derived, assuming that the flow process follows a simplified probabilistic model. The analysis is also extended to the flow size distribution tail. The adequacy of the results to estimate the distributions tails is checked against a real Internet trace.
Keywords :
Internet; statistical distributions; telecommunication traffic; Internet; asymptotic equivalence; flow distributions estimation tails estimation; flow duration distribution tail; flow size distribution tail; probabilistic model; size distributions; Educational institutions; Estimation; Internet; Probabilistic logic; Random variables; Sampling methods; Traffic measurements; flow duration; flow size; heavy-tailed distributions; packet sampling;
Conference_Titel :
Statistical Signal Processing Workshop (SSP), 2012 IEEE
Conference_Location :
Ann Arbor, MI
Print_ISBN :
978-1-4673-0182-4
Electronic_ISBN :
pending
DOI :
10.1109/SSP.2012.6319825