Network topology visualizor using server flow affinity

Foo, Mun Yao (2020) Network topology visualizor using server flow affinity. Final Year Project, UTAR.

Preview

PDF Download (5Mb) | Preview

Abstract

Automatic discovery of network topology is very important and has practical significance due to the fact that network is becoming more and more complex. Technology of computer network is changing and developing rapidly, the importance of scientific and effective network management is becoming increasingly significant. This project is a Network Topology Visualizer for academic purpose. The aim of this paper is improve upon methods and techniques to discover physical and logical topology, then ultimately, incorporating AI to visualize the topology. It will provide readers with the methodology, concept and design Network topology discovery. “Divide and conquer” method will be utilised to segment a physical network for the sake of easing the process of data collection. Then, with the help of machine learning technique known as artificial neural network will be used to help identifying the devices in a topology. The end result of the machine learning is that we are able to obtain an acceptable accuracy result of 75%, although, there is still a lot of room for improvement. The latter part of the project has led to the visualization of the topology data gained from the neural network. The visualizer tool in the project is coded 100% in python, where a JSON configuration file is used to generate the topology of the network to a PDF file. Another JSON files is created in order for the user to see in-depth information on the network. After the development of the visualizer, discussion and the results of the visualizer are made to compare with 2 other network topology visualizer researched online.

Actions (login required)