A State-of-the-Art Survey and Taxonomy for Load Balancing Metrics in SDN Networks

The Software Defined Network (SDN) architecture divides the network into control and data planes. The performance of SDNs is dependent on how traffic is distributed and raises many other challenges. Uneven distribution of the network's load, for instance, can significantly affect the SDN's overall performance. With SDN, forwarding decisions and network management has consolidated in a centralized location. These characteristics enable the development of procedures for controlling traffic using Load Balancing (LB). LB contains different categories such as (Classification, Algorithms, Techniques, and Metrics). In this perspective, we highlight using load balancing metrics (LBMetrics) in SDN networks. This paper provides a comprehensive state-of-the-art survey of LB-Metrics. It gives a thematic taxonomy of LB-Metrics in SDN, including 30 parameters with an explanation for each. Also, a comprehensive study of 41 research papers containing all metrics. This paper aims to know the most important parameters used to measure network performance quality when using load balancing in SDN. Finally, this research collected 30 different metrics and found that most results on SDN balancing use eight metrics as the most critical evaluation that are Throughput (10.6%), Overhead (10.6%), Degree of LB (9.7%), Latency (8.8%), Response Time (8.0%), Packet Loss Rate (6.2%), Resource Utilization (), Transaction Time (4.4%) and Others (37%).