A Machine Learning Methodology for Dynamic QoX Management in Modern Networks

Modern communication networks are increasingly embracing multiple and heterogeneous technologies. Making them work in a cooperative way is crucial to fulfill the user's requirements and enhance their quality of experience (QoE). Additionally, integrative network architectures, such as cloud computing or network virtualization are changing the market and new business opportunities are emerging for service providers. In this complex network environment, many quality indicators (KQI, KPI, KRI, KBO, etc.) are involved. The big challenge is to find the relationships between them to achieve and guarantee the global QoS (QoX). In this paper, a methodology for dynamic QoX management in modern networks is presented. Considering the number of quality indicators and the variety of changing factors that may have an influence on modern network environments, machine learning techniques are proposed to effectively and dynamically manage the QoX.