Network performance evaluation and simulation studies on backbone traffic aggregation

In the modeling of Internet traffic model one of the major challenges faced is limiting the State-Space Explosion problem in description of superposition of MMPP's. Our main aim is to develop: (1) We propose a technique to approximate the traffic aggregation processes described by Markov modulated Poisson processes (MMPP) models, which is based on the clustering structure of the time constants. It is found that the decaying time constants of the aggregated traffic process are the products of the eigenvalues of the transition matrix of the individual traffic. If the time constants are well clustered around some representative time constants (RTC's), the corresponding states can be merged in the state space. (2) After implementing the clustering algorithm, our second aim is to develop an MMPP traffic model and evaluate the performance of the aggregated MMPP traffic stream.; The simulation results show that by aggregating the original traffic stream it is possible to solve the State-Space Explosion problem in description of Superposition of MMPP's and thus we can describe network traffic in a more efficient way. Therefore, the network performance can be evaluated more efficiently.