Open Access

ARTICLE

SIMAD: Secure Intelligent Method for IoT-Fog Environments Attacks Detection

Wided Ben Daoud¹, Sami Mahfoudhi^2,*

1 NTS’Com Research Unit, Sfax, Tunisia
2 Department of Management Information Systems and Production Management, College of Business and Economics, Qassim University, Buraidah, 51452, KSA

* Corresponding Author: Sami Mahfoudhi. Email:

Computers, Materials & Continua 2022, 70(2), 2727-2742. https://doi.org/10.32604/cmc.2022.020141

Received 11 May 2021; Accepted 22 June 2021; Issue published 27 September 2021

Abstract

The Internet of Thing IoT paradigm has emerged in numerous domains and it has achieved an exponential progress. Nevertheless, alongside this advancement, IoT networks are facing an ever-increasing rate of security risks because of the continuous and rapid changes in network environments. In order to overcome these security challenges, the fog system has delivered a powerful environment that provides additional resources for a more improved data security. However, because of the emerging of various breaches, several attacks are ceaselessly emerging in IoT and Fog environment. Consequently, the new emerging applications in IoT-Fog environment still require novel, distributed, and intelligent security models, controls, and decisions. In addition, the ever-evolving hacking techniques and methods and the expanded risks surfaces have demonstrated the importance of attacks detection systems. This proves that even advanced solutions face difficulties in discovering and recognizing these small variations of attacks. In fact, to address the above problems, Artificial Intelligence (AI) methods could be applied on the millions of terabytes of collected information to enhance and optimize the processes of IoT and fog systems. In this respect, this research is designed to adopt a new security scheme supported by an advanced machine learning algorithm to ensure an intelligent distributed attacks detection and a monitoring process that detects malicious attacks and updates threats signature databases in IoT-Fog environments. We evaluated the performance of our distributed approach with the application of certain machine learning mechanisms. The experiments show that the proposed scheme, applied with the Random Forest (RF) is more efficient and provides better accuracy (99.50%), better scalability, and lower false alert rates. In this regard, the distribution character of our method brings about faster detection and better learning.

---

Keywords

---

Citations