Researching of the adequacy of self-similar traffic models used for the quality of service network evaluation

Аuthors

Blagov A. V.

Samara National Research University named after Academician S.P. Korolev, 34, Moskovskoye shosse, Samara, 443086, Russia

e-mail: alex_ssauprof@mail.ru

Abstract

The paper is devoted to the study of telecommunication networks service quality by means of simulation models. Models simulating network traffic with a certain number of parameters at the same time become more and more important in practice due to increased service quality requirements for telecommunication networks. Self-similar network traffic analysis tasks become relevant for its implementation for modeling telecommunication networks during design stage. The paper considers modified telecommunication traffic models «Input M/G/∞» and «On-Off Sources» created by the author. These models allow simulation of self-similar traffic with preset properties and at the same time with predetermined statistical characteristics of small orders. A method is suggested to design and analyze two-dimensional queue distribution in given queueing system where traffic data serves as an input stream. The method is tested with developed modified models of telecommunications traffic: «Input M/G/∞» and «On-Off Sources».
The advantage of the developed model modifications «Input M/G/∞» over the base model has been proved for TCP traffic data. The advantage of developed model «Input M/G/∞» was proved in the course of local pathways simulation. Thus, a confirmation of considered developed models estimation results based on statistic performances analysis? such as the Hurst parameter, mathematical expectation, ACF and univariate distribution was obtained.
Such an approach to two-dimensional queue distribution in given queueing system can be implemented to examine one of the network service quality parameters — delay variation (jitter). Jitter examination and prediction is an important task for telecommunication systems. The obtained results can be used for simulation to analyze the amount of traffic in telecommunication networks in the course of design, as well as during developing methods for bandwidth allocation and reduction of losses in data bursts, for billing systems and other cases where service quality is important.

Keywords:

self-similar traffic, traffic simulation, Input M/G/, On-Off Sources, delay variation, queuing system

References

1. Shelukhin O.I., Tenyakshev A.M., Osin A.V. Fractalnye prosesy v telecommunikaciyakh (Fractal Processes in Telecommunications), Radiotechnica, Moscow, 2003, 480 p.
2. Blagov A.V., Privalov A.Yu. Elektronnyi zhurnal «Trudy MAI», 2010, no 39, available at: http://www.mai.ru/science/trudy/eng/published.php?ID=14802 (accessed 16.08.2010).
3. Mondragon R.J., Arrowsmith D.K., Pitts J.M. Chaotic map for traffic modeling and queuing analysis, Performance Evaluation, 2001, vol. 43, pp.223-240.
4. Gromov Yu.Yu., Zemskoy N.A., Ivanova O.G., Lagutin A.V. Fractalniyi Analiz i prosesy v kompyuternikh setyah (Fractal Analysis and Processes in Computer Networks), Tambov, Technical University, 2007, 65 p.
5. Park K., Willinger W., Wiley Ed. Self-similar Network Traffic and Performance Evaluation, New-York, John Wiley & Sons Inc., 2000, 556 p.
6. Privalov A.Yu., Blagov A.V. Materialy XXIII europeyskoy konferentsii «Modelirovanie i imitatsiya», Madrid, 2009, pp. 18-24.
7. Blagov A.V., Privalov A.Yu. Telecommunicatsii, 2011, no. 6, pp. 7-13.
8. Internet Traffic Archive, http://ita.ee.lbl.gov
9. Blagov A.V. Vestnik Samarskogo gosudarstvennogo aerokosmicheskogo universiteta, 2011, no. 2(26), pp. 201-210.

Download