WITHDRAWN: A hybrid algorithm design for near real time detection cyber attacks from compromised devices to enhance IoT security

Abstract

Abstract With Internet of Things (IoT) technology, the prediction of Nikola Tesla, protean inventor, to realize a “global brain” is becoming a reality with inter-disciplinary innovations. However, constant rise of concerns about IoT security is evident due to tsunami of digitization and connectivity between diversified devices and objects. Wireless connectivity makes the situation worse in consumer IoT security due to lack of vulnerability patching in Information Technology (IT) infrastructure. Traditional security approaches are not directly applicable to IoT use cases as the magnitude of devices, size and computational restrictions. Therefore, there is possibility of many kinds of cyber-attacks such as Denial of Service (DoS), Distributed Denial of Service (DDoS) and so on. Compromised IoT devices may be controlled by botnets or adversaries to launch various attacks. In this paper, we designed a hybrid algorithm known as Advanced Artificial Intelligence (AI) based Cyber Attack Detection (AAI-CAD). It is based on the deep autoencoder and feature extraction integration. UCI dataset is collected for empirical study. The experimental results revealed that the proposed algorithm is efficient to achieve near real-time detection of cyber-attacks in IoT use cases.