Analysis of the Performance of Various Load Balancing Algorithms for SDN in Data Center Networks

Abstract

In order to fulfil the demands of modern IT, it is necessary to address the many limitations of current networking infrastructures. Software Defined Network (SDN) is the new networking technique to get around these constraints. Existing Network Infrastructures use static switches, which results in a substandard utilization of the network resources, which is one of the main problems. The main issues with the present networking trend are the slow response and delays. In this research different load-balancing algorithms for Software Defined Network based Data Center network is implemented to overcome mentioned issues. The POX controller is used in the implementation and the Mininet emulator is used to simulate the network functionality. The load balancing algorithms program is written in the Python language, which is also used for creation of the network topology. Finally, Siege HTTP load testing tool is used to test network performance. The testing is focused on Quality of Service pa- rameters transaction rate, average response time and throughput. The results show that the Weighted algorithm outperforms all other considered algorithms.