IoT Network Traffic Classification and attack detection based on Network Traffic Characteristics using Artificial Intelligence

Introduction

A wide range of embedded devices apply to the internet of Things (IoT) internet connected, allowing them to send and exchange information in intelligent environments for one another. Since these IoT devices transmits their network traffic in broadcast mode due to wireless media, it is simple for an intruder to collect data by analyzing the network traffic of IoT devices. In addition, malicious network traffic can be generated by a malicious IoT devices that other IoT devices can be corrupted, Denial of Service (DoS) attacks can be initiated, installing using malware etc. Therefore, it is necessary for an administrator to evaluate the network traffic for security management and allocation of resources.

Objective

• Build a coherent IoT traffic classification structure. We will extract different IoT network traffic functionality and illustrate the value of each attribute for the classification.
• Using a set of machine learning algorithms, each classifier can compare the obtained prediction performance based on precision, training and accuracy, algorithm time, error rate, etc.
• Develop an artificial model to define the pattern of attack by tracing the IoT devices traffic.

Resources

Outcomes

• Patent Published: "NETWORK DEVICE CLASSIFIER FOR INTERNET OF THINGS (IOT) NETWORKS".
  Inventor: G. Singal, Mayank Swarnkar, Rakesh Kumar, Manish Sharma, Vinayak Joshi, Suneet K Gupta, Deepak Gagr, Publication Date: December 12, 2021, Application Number: 202111055728.
• Paper Published: R. Kumar, M. Swarnkar, Gaurav Singal, N. Kumar. "IoT Network Traffic Classification using Machine Learning Algorithms: An Experimental Analysis.", Internet of Things, IEEE, October 2021.