UDC 621.391
COMPARATIVE ANALYSIS OF NETWORK TRAFFIC DECORRELATION METHODS
V. A. Osanov, post-graduate student, PSUTI, Samara, Russia;
orcid.org/0000-0001-6031-3306, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
Based on experimental data confirming the relationship between network traffic autocorrelation and video frame losses and an increase in average packet transmission time, a method for solving the problem of improving service quality by decorrelation is proposed. The aim of this work is to select an efficient method for reducing the autocorrelation level to perform decorrelation process of time intervals sequence between packets in network based on their comparative assessment. Main criteria for comparison, of which the prior ity is decorrelation process execution time with the least impact on network delay, are formulated. Known methods such as Karhunen – Loève decomposition, discrete cosine transform, and others, do not satisfy this criterion and, as a result, cannot be used for practical implementation of decorrelation process. Methods based on Haar wavelet transform and permutation are proposed. Their comparison is performed by software implementation of the methods, analysis of decorrelation results of simulated autocorrelation sequence, and calculation of computational complexity of execution algorithms for different input data parameters. The obtained results confirm the advantage of proposed methods and, in particular, spiral permutation method.
Key words: : network traffic, correlation traffic, quality of service, decorrelation, Karhunen – Loève de composition, discrete cosine transform, Haar wavelet transform, S-random permutation, spiral permutation.