Managementul resurselor în soluții de virtualizare pentru rețele și servicii / Resource management in virtualization solutions for networks and services

This thesis aims to present the NFV and Network Slicing paradigms and their various use cases. The study of two NFV frameworks such as SONATA and OSM complements the study of these paradigms. For a better understanding of how these frameworks work, various experiments have been conducted. After analyzing the results obtained from these experiments, one out of these two frameworks has been chosen to be integrated into a standalone system allowing intelligent configuration of new network slices using Machine Learning (ML) algorithms. The autonomous system was integrated into a platform through which a new service chain was built, allowing experiments to be carried out to validate this implementation. This autonomous system and the platform that includes it are original contributions of the author, both at the architectural, implementation and test stages. By using this automated platform, deploying a network slice through the Open Source MANO - Virtual Infrastructure Manager service chain becomes faster and easier. Adding the intelligent configuration block enables network slices to be optimized both in terms of deployment and installation on the appropriate infrastructure, as well as adaptive change of alarm thresholds and scalability. The solution proposed in this paper aims to improve the current network slice management workflow through automated mechanisms that can optimally allocate resources and adapt alarm and scaling thresholds to optimize network slice operations