A Novel approach for Designing a Blockchain-Based Intrusion Detection System for Securing IoT Networks Using Machine Learning Algorithms

Main Article Content

Amit Saxena, Ravula Arun Kumar, Kodge B. G., Rajesh Gadipuuri, R.Menaka, Chaithrashree.A, Dr. Avinash

Abstract

  The Internet of Things (IoT) explosion has placed huge numbers of connected devices in operational environments, and that fragility leaves these networks wide open for attackers. In order to catch possible attacks effectively, it is always necessary to have Intrusion Detection Systems (IDS), but 80Performing IDS on IoT Networks by Siddharth Sridhar types of IDS approach where they are dealt with Traditional IDS methods scale really low and inaccurate in IoT networks. Available solutions generally adopt the signature-based or anomaly-based detection methods, however they are not effective for unknown threats crisism and lack of real-time responsiveness. In this paper, we provide a novel solution of securing reinforcement learning for Internet of Things Networks on Blockchain education. Utilizing the decentralization of blockchain, this system securely deposits and checks IDS logs in a way that is reliable across an entire network. This allows us to create a mathematical model integrating Random Forest and K-Means Clustering methods for the classification of threats as well as anomaly detection. This methodology is proposed with the aim to have low latency and high accuracy when detecting malicious activities. Experimental results also show that the system achieves over 95% detection rate as well as more effective reduction of false-positive rates than traditional Intrusion Detection System (IDS) solutions. Furthermore, by incorporating blockchain technology, the system benefits from strong security guarantees and becomes a more tamper-evident and immune-to-a-3rd-party-approach. This approach is useful in areas such as smart homes, industrial IoT systems, and critical infrastructure protection where real-time threat detection is of utmost importance. Nevertheless, an underlying computational overhead and scalability issue persists owing to the limited resource availability in IoT devices and complexity involved in blockchain transactions. Further research needs to be incorporated in optimizing these categories of study to make them widely adopted and efficient in deploying at scale in IoT environment.

Article Details

Section
Articles