The questions of asymptotic research of turns are considered in the high-speed networks of telecommunications and the scopes of the offered approximations of models are certain. Researches, allowing to define necessary exactness and calculable power, are conducted, which are required for the calculation of approximation of maximum. It is set that exponential temporal scales are optimum for the traffic of fractal brownian motion in sense of concordance. The conducted experiments showed influencing of tails of distributing in different time scale factors on the process of discipline of queues. Observed, that at nonGaussian traffic scenarios a correlation structure (short-term and long-term) describes the conduct of turns it is not enough adequately. Directions of further researches are certain in this connection.
telecommunication network, asymptotic analysis, non-Gaussian process, fractal brownian motion, fractal Gaussian noise, multifractal veivlet-model