Abstract:
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%.
