Estimating QoS parameters in an SDN node when processing correlated flows

The ever-increasing volume of transmitted traffic has led to the development of new infocommunication technologies designed to improve the quality of service (QoS) of network nodes. The use of optical networks is crucial in addressing the challenges of ensuring quality of service. However, managing of optical networks to provide differentiated services for users is a complex process that demands innovative approaches and technologies. One such technology is software-defined networks (SDN), which involves separating the routing plane and switching of traffic flows. Analyzing the performance of any network node should take into account the specific design features of this separation and traffic correlation properties. The paper analyzes the effect of the correlation of time intervals between incoming packets on the performance of a node in a software-defined network operating on the Open Flow protocol. Analytical expressions are derived for calculating the average values of packet delay time in a network node and its variation. In this case, the analytical model of a network node is represented by the G/G/1 system, where the probability densities of time intervals between incoming packets and packet processing time intervals are specified by hyperexponential models. This approach is particularly suitable for ensuring quality of service in optical networks, which are controlled using SDN.