Inferring the background traffic arrival process in the Internet

Péter Hága, I. Csabai, G. Vattay

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Phase transition has been found in many complex interactivity systems. Complex networks are not exception either but there are quite few real systems where we can directly understand the emergence of this nontrivial behavior from the microscopic view. In this paper, we present the emergence of the phase transition between the congested and uncongested phases of a network link. We demonstrate a method to infer the background traffic arrival process, which is one of the key state parameters of the Internet traffic. The traffic arrival process in the Internet has been investigated in several studies, since the recognition of its self-similar nature. The statistical properties of the traffic arrival process are very important since they are fundamental in modeling the dynamical behavior. Here, we demonstrate how the widely used packet train technique can be used to determine the main properties of the traffic arrival process. We show that the packet train dispersion is sensitive to the congestion on the network path. We introduce the packet train stretch as an order parameter to describe the phase transition between the congested and uncongested phases of the bottleneck link in the path. We find that the distribution of the background traffic arrival process can be determined from the average packet train dispersion at the critical point of the system.

Original languageEnglish
Article number066103
JournalPhysical Review E - Statistical, Nonlinear, and Soft Matter Physics
Volume80
Issue number6
DOIs
Publication statusPublished - Dec 3 2009

Fingerprint

traffic
arrivals
Traffic
Phase Transition
Internet Traffic
Path
Interactivity
Stretch
congestion
Dynamical Behavior
Congestion
Complex Networks
Order Parameter
Statistical property
Demonstrate
Exception
Critical point
complex systems
Background
critical point

ASJC Scopus subject areas

  • Condensed Matter Physics
  • Statistical and Nonlinear Physics
  • Statistics and Probability

Cite this

Inferring the background traffic arrival process in the Internet. / Hága, Péter; Csabai, I.; Vattay, G.

In: Physical Review E - Statistical, Nonlinear, and Soft Matter Physics, Vol. 80, No. 6, 066103, 03.12.2009.

Research output: Contribution to journalArticle

@article{3d56188725db4882892667584e5788ea,
title = "Inferring the background traffic arrival process in the Internet",
abstract = "Phase transition has been found in many complex interactivity systems. Complex networks are not exception either but there are quite few real systems where we can directly understand the emergence of this nontrivial behavior from the microscopic view. In this paper, we present the emergence of the phase transition between the congested and uncongested phases of a network link. We demonstrate a method to infer the background traffic arrival process, which is one of the key state parameters of the Internet traffic. The traffic arrival process in the Internet has been investigated in several studies, since the recognition of its self-similar nature. The statistical properties of the traffic arrival process are very important since they are fundamental in modeling the dynamical behavior. Here, we demonstrate how the widely used packet train technique can be used to determine the main properties of the traffic arrival process. We show that the packet train dispersion is sensitive to the congestion on the network path. We introduce the packet train stretch as an order parameter to describe the phase transition between the congested and uncongested phases of the bottleneck link in the path. We find that the distribution of the background traffic arrival process can be determined from the average packet train dispersion at the critical point of the system.",
author = "P{\'e}ter H{\'a}ga and I. Csabai and G. Vattay",
year = "2009",
month = "12",
day = "3",
doi = "10.1103/PhysRevE.80.066103",
language = "English",
volume = "80",
journal = "Physical review. E",
issn = "2470-0045",
publisher = "American Physical Society",
number = "6",

}

TY - JOUR

T1 - Inferring the background traffic arrival process in the Internet

AU - Hága, Péter

AU - Csabai, I.

AU - Vattay, G.

PY - 2009/12/3

Y1 - 2009/12/3

N2 - Phase transition has been found in many complex interactivity systems. Complex networks are not exception either but there are quite few real systems where we can directly understand the emergence of this nontrivial behavior from the microscopic view. In this paper, we present the emergence of the phase transition between the congested and uncongested phases of a network link. We demonstrate a method to infer the background traffic arrival process, which is one of the key state parameters of the Internet traffic. The traffic arrival process in the Internet has been investigated in several studies, since the recognition of its self-similar nature. The statistical properties of the traffic arrival process are very important since they are fundamental in modeling the dynamical behavior. Here, we demonstrate how the widely used packet train technique can be used to determine the main properties of the traffic arrival process. We show that the packet train dispersion is sensitive to the congestion on the network path. We introduce the packet train stretch as an order parameter to describe the phase transition between the congested and uncongested phases of the bottleneck link in the path. We find that the distribution of the background traffic arrival process can be determined from the average packet train dispersion at the critical point of the system.

AB - Phase transition has been found in many complex interactivity systems. Complex networks are not exception either but there are quite few real systems where we can directly understand the emergence of this nontrivial behavior from the microscopic view. In this paper, we present the emergence of the phase transition between the congested and uncongested phases of a network link. We demonstrate a method to infer the background traffic arrival process, which is one of the key state parameters of the Internet traffic. The traffic arrival process in the Internet has been investigated in several studies, since the recognition of its self-similar nature. The statistical properties of the traffic arrival process are very important since they are fundamental in modeling the dynamical behavior. Here, we demonstrate how the widely used packet train technique can be used to determine the main properties of the traffic arrival process. We show that the packet train dispersion is sensitive to the congestion on the network path. We introduce the packet train stretch as an order parameter to describe the phase transition between the congested and uncongested phases of the bottleneck link in the path. We find that the distribution of the background traffic arrival process can be determined from the average packet train dispersion at the critical point of the system.

UR - http://www.scopus.com/inward/record.url?scp=72449173371&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=72449173371&partnerID=8YFLogxK

U2 - 10.1103/PhysRevE.80.066103

DO - 10.1103/PhysRevE.80.066103

M3 - Article

C2 - 20365227

AN - SCOPUS:72449173371

VL - 80

JO - Physical review. E

JF - Physical review. E

SN - 2470-0045

IS - 6

M1 - 066103

ER -