Uma Nova Abordagem para Estimação da Banda Efetiva em Processos Fractais (A new Bandwidth Estimation Approach for Fractal Processes)

Firmiano Perlingeiro (firmiano@decom.fee.unicamp.br), Lee-Luan Ling (lee@decom.fee.unicamp.br)


Departamento de Comunicacoes (DECOM), Faculdade de Engenharia Eletrica e Computacao (FEEC), Universidade Estadual Campinas (UNICAMP) - SP, Brasil.
This paper appears in: Revista IEEE América Latina

Publication Date: Dec. 2005
Volume: 3,   Issue: 5 
ISSN: 1548-0992


Abstract:
Recent extensive traffic analyses carried out over different network technologies have shown network traffic in its "monofractal" or "multifractal" nature as well as in its impact on network performance. Based on the results of these "monofractal" and "multifractal" traffic studies we introduce a new traffic parameter, capable of simultaneously expressing both "monofractal" and "multifractal" characteristics. This traffic characterization parameter is used for the design of more accurate effective bandwidth estimation approaches. In order to achieve a general and good methodology for effective bandwidth estimation for given traffic when the QoS requirements is take into account, we propose a broader optimization procedure. For the validation of the proposed effective bandwidth estimation approach, we intensely performed tests using real traffic traces and simulations. The obtained analytical and experimental results clearly indicate that the proposed bandwidth estimation approach for fractal traffic processes satisfies the QoS requirements for all analyzed traffic traces.

Index Terms:
Effective bandwidth, buffer, delay, overflow probability, quality of service (QoS), self-similar processes, long range dependence, "multifractal", Markov processes, traffic parameters.   


Documents that cite this document
This function is not implemented yet.


[PDF Full-Text (439)]