engleski

Network Performance Evaluation by Means of Self Similar Traffic Model

The paper focuses on a difference in buffer size dimensioning concerning different input traffic types. Poisson traffic, burst traffic with exponentially distributed burst lengths and self- similar traffic were generated. Mean and maximum queue lengths were calculated by varying the input traffic load. The time variations of the packet number in the queue were also analyzed. Self- similarity was accomplished using a Pareto traffic aggregation and fractional Gaussian noise. Calculations were based on the simulation results from the object-oriented model and discrete event simulation in the Cosmos tool.

network performance; self similar traffic

nije evidentirano