Machine Learning Driven Load Balancing in Software Defined Networks: Recent Progress and Emerging Challenges

Load balancing has arisen as an important challenge in case of Software Defined Network because of the dynamic patterns in the traffic and expanding demands of network. The centralized framework of SDN delivers unique opportunity. It enhances the performance of network and usage of resources by distributing the load of network dynamically across multiple devices and paths. This paper provides a concise overview of SDN and examines the influence of load balancing within this context. It includes a comprehensive study to analyse the algorithms of load balancing used by researchers for different scenarios and to identify the persistent challenges. Furthermore, a methodology has been proposed to highlight the impact of machine learning techniques to amplify the load balancing decision which can be useful to develop more intelligent and adaptive architectures of SDN. For future research, the study aims to deploy scalable and robust SDN solutions in next-generation network infrastructures.