A note on simulation of LRD network traffic
Ming Li, Wei Zhao · 2009
Abstract:- This paper addresses the generation of long-range dependent (LRD) network traffic with very big variance. In doing so, it first reviews two commonly used methods of random data generation. The one is the random data generation according to a given probability density function (PDF). The other is the method of synthesizing random data based on a given power spectrum density (PSD) function. Then, the limitations of both methods in synthesizing long-range dependent (LRD) network traffic with very big variance are explained. Based on those, the method of random data generation for a predetermined autocorrelation function (ACF) is discussed. Its advantages in comparison with the PDF-based method and the PSD-based one are interpreted. In addition, three methods of the white noise generation are addressed. Key-Words:- Random data generation; Long-range dependence; 1/f noise; Network traffic.