Bivariate statistical analysis of TCP-flow sizes and durations

Natalia Markovich; Jorma Kilpi

Bivariate statistical analysis of TCP-flow sizes and durations

Natalia Markovich, Jorma Kilpi

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific

Abstract

This paper is devoted to the nonparametric analysis of heavy-tailed bivariate data. The TCPflow sizes and durations measured from the mobile network of the Finnish operator Elisa are investigated. The rough methods to detect heaviness of tails from [4] and [8] are applied and it is shown that both flow sizes and durations are heavy-tailed. An extreme value bivariate distribution of two random variables, TCP flow sizes and durations, is estimated by nonparametric methods. For this purpose, the Pickands dependence function is estimated from the empirical sample of the observed pairs. Furthermore, the respective bivariate quantile plot is constructed.

Original language	English
Title of host publication	Proceedings of the EuroNGI Workshop on Stochastic Performance Models for Resource Allocation in Communication Systems, StoPeRa'06
Pages	47 - 50
Publication status	Published - 2006
MoE publication type	B3 Non-refereed article in conference proceedings
Event	Stochastic Performance Models for Resource Allocation in Communication Systems, StoPeRa - Amsterdam, Netherlands Duration: 8 Nov 2006 → 10 Nov 2006

Conference

Conference	Stochastic Performance Models for Resource Allocation in Communication Systems, StoPeRa
Country/Territory	Netherlands
City	Amsterdam
Period	8/11/06 → 10/11/06

Cite this

@inproceedings{77f06e41f300449d913ae47773418a1e,

title = "Bivariate statistical analysis of TCP-flow sizes and durations",

abstract = "This paper is devoted to the nonparametric analysis of heavy-tailed bivariate data. The TCPflow sizes and durations measured from the mobile network of the Finnish operator Elisa are investigated. The rough methods to detect heaviness of tails from [4] and [8] are applied and it is shown that both flow sizes and durations are heavy-tailed. An extreme value bivariate distribution of two random variables, TCP flow sizes and durations, is estimated by nonparametric methods. For this purpose, the Pickands dependence function is estimated from the empirical sample of the observed pairs. Furthermore, the respective bivariate quantile plot is constructed.",

author = "Natalia Markovich and Jorma Kilpi",

note = "Project code: 6374; Stochastic Performance Models for Resource Allocation in Communication Systems, StoPeRa ; Conference date: 08-11-2006 Through 10-11-2006",

year = "2006",

language = "English",

pages = "47 -- 50",

booktitle = "Proceedings of the EuroNGI Workshop on Stochastic Performance Models for Resource Allocation in Communication Systems, StoPeRa'06",

}

TY - GEN

T1 - Bivariate statistical analysis of TCP-flow sizes and durations

AU - Markovich, Natalia

AU - Kilpi, Jorma

N1 - Project code: 6374

PY - 2006

Y1 - 2006

N2 - This paper is devoted to the nonparametric analysis of heavy-tailed bivariate data. The TCPflow sizes and durations measured from the mobile network of the Finnish operator Elisa are investigated. The rough methods to detect heaviness of tails from [4] and [8] are applied and it is shown that both flow sizes and durations are heavy-tailed. An extreme value bivariate distribution of two random variables, TCP flow sizes and durations, is estimated by nonparametric methods. For this purpose, the Pickands dependence function is estimated from the empirical sample of the observed pairs. Furthermore, the respective bivariate quantile plot is constructed.

AB - This paper is devoted to the nonparametric analysis of heavy-tailed bivariate data. The TCPflow sizes and durations measured from the mobile network of the Finnish operator Elisa are investigated. The rough methods to detect heaviness of tails from [4] and [8] are applied and it is shown that both flow sizes and durations are heavy-tailed. An extreme value bivariate distribution of two random variables, TCP flow sizes and durations, is estimated by nonparametric methods. For this purpose, the Pickands dependence function is estimated from the empirical sample of the observed pairs. Furthermore, the respective bivariate quantile plot is constructed.

M3 - Conference article in proceedings

SP - 47

EP - 50

BT - Proceedings of the EuroNGI Workshop on Stochastic Performance Models for Resource Allocation in Communication Systems, StoPeRa'06

T2 - Stochastic Performance Models for Resource Allocation in Communication Systems, StoPeRa

Y2 - 8 November 2006 through 10 November 2006

ER -

Bivariate statistical analysis of TCP-flow sizes and durations

Abstract

Conference

Fingerprint

Cite this