Tensor Based Load Balancing under Self-Similar Traffic Properties with Guaranteed QoS

Abstract

The paper proposes a tensor based load balancing multiflow model under self-similar traffic properties with guaranteed QoS. The novelty of the proposed model is that it is aimed at providing demanded values of QoS indicators such as flow rate and average packet delay taking into account the self-similarity of data traffic when calculating average packet delay in the link, which was also used in formation of the metric tensor. Providing the required numerical values of the QoS indicators was achieved by introducing into the model the corresponding Quality of Service ensuring conditions. The numerical research results for the data traffic flows with different values of QoS indicators confirmed the adequacy of the obtained solutions.