An Investigation on Fat Tree Topology ARP Spoof Detection Using Software Defined Networking

Using SDN's benefits, the authors of this paper propose a novel load balancing algorithm for use in data centre networks. For the purposes of simulating and evaluating the proposed design, Mininet was used as an emulation tool, while Miniedit was used as a graphical user interface tool. Fat-Tree topology with the following parameters was used to achieve a realistic setting for the data centre network: 5 pods, 17 edge switches, 17 aggregation switches, 4 core switches, and 16 hosts. In order to simulate real-world traffic as closely as possible, a variety of scenarios and traffic distributions were used. As an SDN controller, POX was selected. With better throughput and lower loss rates across the board, the proposed design is a clear winner over the status quo. In the first case, new hosts are added, while in the second case, the existing connections experience a surge in demand.

I. Introduction

Information may be sent from individual place to another in the form of mathematical packets on account of the delivered control and transport network pacts operating inside the routers and switches. While IP networks have enhanced ever-present, the complicatedness and trouble of directing bureaucracy has restricted their use. Network administrators must manually construct each network scheme accompanying reduced-level; frequently hawker-particular commands so that express the asked high-ranking network procedures. Network atmospheres again should handle the changeable character of sins and changes in load, in addition to the complicatedness received for one arrangement itself. Unfortunately, means for mechanical reconfiguration and answer are predominantly missing in contemporary's IP networks. Therefore, it is intensely troublesome to apply the essential tactics in aforementioned an active scene [1].