Network traffic prediction

Abstract

In order for businesses to maintain their productivity today, it iscrucial that they constantly monitor their networks to avoid potential threats such as network failures.It has become a necessity for businesses of all sizes to employ traffic monitoring tools. Traditional network traffic monitoring tools such as Netflow are restricted to collecting traffic across the Internet from routers. As networks evolve, networks can be further divided into smaller partitions called Virtual LANs.As such, there is a need to gain visibility into network information at a lower level. The purpose of this project is to understand how traffic monitoring tools work and how they are able to provide a more reliable network by analyzing the information that they provide. Using a technology called sFlow, we can obtain more information about traffic flow. By using different data visualization methods, we can analyze and draw insights from the data presented in each method. Subsequently, I implemented a web application that allows the administrators to upload traffic data and display the data in graphs. Different data visualization methods can display different aspects of the network. Organizations should make use of various data visualization tools to present a more complete view of their network. Possible future work include creating baselines to determine normal or “healthy” levels of traffic flow.