An Envelope Process for Farima Traffic Model

The paper presents an envelope process which is a deterministic function bounding a stochastic process for a FARIMA traffic model. The envelope process is defined by FARIMA parameters (d associated with Hurst parameter, ϕ or θ) and considered exceed probability ε , which determines the accuracy of the envelope process. The results obtained allow for analytical modeling of network behaviour fed by a FARIMA traffic. 1. INTRODUCTION A Fractional Autoregressive Integrated Moving Average (FARIMA) process is widely used in video traffic modeling [5, 8]. Modern mobile networks need precise analysis because video on demand applications are becoming more and more popular. A self-similarity (SS) of network traffic makes network analysis much more difficult. An envelope process permits changing stochastic equations to analytical ones and in this way helps greatly to model related queuing phenomena.