Development and Testing of Traffic Balancing Techniques in SDN Networks Based on the Combined QoS Parameter

Abstract

Actuality. The growth of real-time traffic places increased demands on the quality of service in networks. Traditional load balancing methods, focused only on network node loading, do not take packet loss, delay, and jitter into account. There is a need for advanced balancing algorithms for software-configurable networks capable of dynamically taking into account both the traffic load on network nodes and its quality characteristics.

Purpose: development, theoretical justification, and experimental testing of network traffic balancing techniques based on a combined quality parameter for software-configurable networks, ensuring compliance with client traffic service quality requirements.

Methods: theocritical analysis of SLA communication operators to identify key parameters of the quality of service to all networks, mathematical modeling and formalization of the normalization procedure of various quality parameters, traffic balancing method based on the modified Enhanced Weighted Round-Robin algorithm, with the use of developed software for simulating traffic balancing in SDN networks. The solution. Development

analysis of dynamic balancing techniques, with integrated calculation of the combined service quality parameter in networks. A mechanism for normalization of quality parameters is proposed.

Experimentally proven effectiveness: the algorithm ensured uniform distribution of the load and improved the combined indicator of the quality of service in networks by 46% compared to the method of load balancing only.

Novelty. For the first time, an original adaptive model of mathematical normalization of various Quality of Service parameters was developed. Specialized software was created for simulation modeling and visualization of the balancing method taking into account Quality of Service.

Practical significance. Development of a ready-made tool for modeling and testing infrastructure stability. Demonstration of practical implementation of intellectual balancing for implementation in SDN controllers.