Implementation and Performance Evaluation of Load Balanced Routing in SDN based Fat Tree Data Center

As the number of data intensive and distributed applications increase, data centers are highly utilized for their storage, processing and accessing. These applications cause throughput sensitive incast traffic in the data center. Traditional shortest path routing algorithms cause traffic hotspots in the shortest path links and lead to congestion. Therefore, load balanced routing strategies need to be employed to spread the traffic across multiple paths. This work employs the idea of multi path routing to achieve load balancing in SDN based Fat tree Data center. It implements Equal Cost Multi Path routing algorithm on Fat tree Data Center Network in Ryu SDN controller and evaluates its load balancing performance against traditional Shortest Path Routing algorithm under different data center benchmark traffic scenarios. We have conducted an exhaustive study about the load balancing effect of ECMP on Ryu based SDN Fat tree data center and found that ECMP increases Average network throughput by 68% and Average Link utilization by 60%. Also packet drop is decreased by 44% with ECMP.