ABSTRACT
The Internet of Things (IoT) has become the biggest sector for commercialization as well as industrialization in this era. Automation in the machinery, machine-to-machine communication, and self-sustainable infrastructures are the baseline for the evolution of industrial IoT. The underlying architecture of IoT favors’ expansion of the Internet of Vehicles (IoV) and other Internet-based technology evolutions. As the technology expands, it should be able to provide the platform for the upcoming trends in that domain thereby enhancing the adaptability. Interoperability and compatibility are the two important characteristics of IoT. When there is an enhancement to the technology, it’s obvious that there will be a space for a security breach. The number of devices that are connected in the IoT will be exponential, where the relationship between them is completely dependent on the connectivity. By using that as a loophole, the intruders/attackers try to compromise the target thereby breaching the network security. In this chapter, we have explored various security threats that are common in IoT and how deep learning (DL) can be used to mitigate the risks involved with them. We have compared various algorithms that are proposed in the recent past to mitigate the threats by involving deep/machine learning models, and we propose a strategy that integrates DL algorithms for attack detection and mitigation. Some of the security threats we will be addressing in this chapter are Buffer reservation attack, Rouge Server, Jamming, DoS (Denial of Service), etc.
