A Minimum Spanning Tree-based Routing Technique of FAT Tree for Efficient Data Center Networking

Abstract

Data centers (DCs) have become the critical and irreplaceable infrastructure to drive this ever-growing trend because of the exponential expansion of Internet services. In general, DC has several data & storage nodes connected by a network that is specially built, called Data Center Network (DCN). DCN functions as a communications backbone also play a significant role in maximizing the operation of DC. However, the infrastructure and operations are very difficult in comparison with the conventional network compared with unique DCN specifications, including a wide range of applications, high power density, and high reliability. In this paper, the Bellmanford MST algorithm is used for FAT-tree-based routing. The main characteristic of the FAT Tree is that the links which are used for connecting nodes from different levels may have different bandwidth. The Bellmanford algorithm used here is the shortest path algorithm which provides a reliable and short path for the route nodes to travel which provides fewer time complexities in data center networking.