A Dynamic Algorithm for Optimization of Network Traffic through Smart Network Switch Data Flow Management

Abstract

Networking affects our daily interaction with the digital world. Advancements in Smart Device technology have led to the need for faster, more power-efficient, and error-free networks. Networking protocols, such as TCP and UDP, were originally designed without these new requirements and now face significant challenges in adapting to those goals i.e., efficiently handling congestion delays in switching networks, increasing production costs, and higher packet losses. In this paper, we propose the Smart Switch Dynamic Delay Algorithm (SSDDA) which takes into consideration multiple factors including link utilization, application-level throughput (goodput), packet loss, and queuing delay of the TCP/UDP link. The application is in datacenter switches where the switching could be moved from centralized controller to decentralized data path and programmable smart switches. Continuous feedback from the network enables SSDDA to estimate delay and modify packet flows for optimal performance. Simulation results using NS3 in a fat tree topology are carried out validating the algorithmic benefits of SSDDA.