Selection of Topologies in Ryu Controller-Based Data Center for Handling IoT Data

A Data Center Network (DCN) is a centralized place where servers, computer systems, networking hardware, and storage devices are housed to store, process, manage, and distribute large amounts of data. The Internet-of-Things (IoT) devices with multiple sensors deployed throughout the environment collect data from their surroundings. Once this data is collected, it must be stored securely and efficiently in the data center. The topology of a data center refers to the physical and logical arrangement or structure of the network and computer resources of the data center. This paper is ideal for comparing the performance of the topologies, namely, Fat-tree, BCube, and DCell. In the experiments, a popular SDN emulator, Mininet, obtains the best output possible for throughput, jitter, and latency parameters. Valuation is performed by the controller RYU, and the results clearly show that BCube achieves better throughput. Fat-tree performs better in jitter, and the DCell topology is the most effective for reachability.