An Approximate Performance of Self-Similar Lognormal M 1 K Internet Traffic Model

Abstract

Modeling Internet traffic data with the inherent condition of concurrent arrival of packets requires the use of heavy-tailed distributions. In this paper, we present log-normal distribution as a suitable heavy-tailed distribution for modeling self-similar Internet arrival process. Specifically, we developed its approximate performance measures for large buffer size. Results using the simulated data confirm the adequacy of the proposed model.