Traditional networking architectures have many significant limitations that must be overcome to meet modern IT requirements. To overcome these limitations; Software Defined Networking (SDN) is taking place as the new networking approach. One of the major issues of traditional networks is that they use static switches that cause poor utilization of the network resources. Another issue is the packet loss and delay in case of switch breakdown. This research proposes an implementation of a dynamic load balancing algorithm for SDN based data center network to overcome these issues. A test-bed has been implemented using Mininet software to emulate the network, and OpenDaylight platform (ODL) as SDN controller. Python programming language is used to define a fat-tree network topology and to write the load balancing algorithm program. Finally, iPerf is used to test network performance. The network was tested before and after running the load balancing algorithm. The testing focused on some of Quality of Service (QoS) parameters such as throughput, bandwidth, delay, jitter, and packet loss between two servers in the fat-tree network. The algorithm increased throughput with at least 32.3%, and improved network utilization. However, in large networks it increased mean jitter from 0.3736 ms to 3.2891 ms, and it increased packet loss by 4.9%.