Packet Loss Probability Dependence on Number of ON-OFF Traffic Sources in OPNET

Abstract

For the last years the characteristics of the telecommunication networks traffic have changed. It happened due to spreading of multimedia-based services. Many research works show that nowadays traffic has self-similar behavior. The traditional queuing models do not take into account the long-range dependence of the self-similar traffic. This raises an issue of the development of the efficient MBAC model that would be adequate for the real traffic. The efficiency of the MBAC model was evaluated according to overflow probability. The paper presents results gained during the evaluation of the queuing model with the input traffic with self-similar characteristics. For the estimation of the model the OPNET simulation software was used. The self-similar traffic was generated using ON-OFF model. The results indicated that buffer overflow probability depends on ON-OFF traffic sources. Implications of the current study and directions for future research are discussed. Ill. 6, bibl. 10 (in English, summaries in English, Russian and Lithuanian).