A Queuing Model for Software-Defined Networking in Cloud and Multi-Controller Environments
Islem Teboub, Amar Aïssani, Fatiha Merazka · 2024
Enterprises are increasingly adopting Software-Defined Networking (SDN) for advanced infrastructure management, cost reduction, and adaptation to technological challenges. Cloud-based SDN controllers play a pivotal role in ensuring flexibility, security, and scalability. The transition from a single-controller to a multi-controller paradigm in cloud networks not only improves security and availability but also reinforces the overall resilience of SDN. Our scientific study aims to propose a novel approach for deploying SDN networks and to introduce a queueing model for this proposed concept. The work focuses on modeling SDN networks using Markovian stochastic processes, taking into consideration the availability criterion through the multiplicity of SDN controllers at the control layer, specifically utilizing queuing theory. The proposed model brings improvement and optimization compared to existing ones. The evaluation of Quality of Service (QoS), based on response time measurements, has demonstrated the efficiency of the model. Additionally, observing the model through the conducted experiments has led us to conclude that scalability is ensured concerning the multiplicity of controllers. Our model has successfully addressed the multi-criteria constraint and ensured an optimal combination that incorporates cloud concepts, QoS, and multi-controllers.