No abstract

The article aims to develop a model for forecasting the characteristics of traffic flows in real-time based on the classification of applications using machine learning methods to ensure the quality of service. It is shown that the model can forecast the mean rate and frequency of packet arrival for the entire flow of each class separately. The prediction is based on information about the previous flows of this class and the first 15 packets of the active flow. Thus, the Random Forest Regression method reduces the prediction error by approximately 1.5 times compared to the standard mean estimate for transmitted packets issued at the switch interface.

We consider a data transmitting system with an active queue management designed to prevent overloading, where fuzzy logic controller is used. We developed a mathematical model that takes into account the features of the data transfer system with an active queue management, which keeps the queue length in the range of values close to a given reference value of the queue length. The method of hysteretic control for incoming load with two thresholds was used as a basis of the model.The mathematical model is a queuing system with a threshold control, which is designed for the analysis of the possibility of hysteresis in modeling of systems with active queue management. The model was described by a Markov process, for which the numerical solution of the equilibrium equations was obtained, steady state probabilities were calculated. The main probabilistic measures are the following: the mean value and the standard deviation of a queue length, and the probability for the queue length of being within the specified limits from the reference value. The numerical analysis in the load range, which includes a system overload, indicated the adequacy of the constructed mathematical model with hysteretic control and system with an active queue management based on fuzzy logic controller. The proposed fuzzy logic method was implemented for Linux kernel and the test results show better quality of service parameters than other tested methods.

No abstract

No abstract