In this paper deals with the Internet Security, today's IT leaders face numerous problems and quick developments. They must safeguard corporate, consumer, citizen, member, and employee data while fending off cyber-attacks. Intrusion Detection System (IDS) is a mature technology architecture that is primarily designed to safeguard the network from external cyber-attacks. With the growth of the Internet and the evolution of cyber-attacks, it is more vital than ever to build new cyber security tools, especially for Internet of Things (IoT) networks. This study presents a thorough examination of the use of deep learning (DL) technologies in cyber security. After that, we show how learning differs from deep learning. Furthermore, a description of current cyber-attacks in IoT and other networks, as well as the efficacy of DL approaches for managing these attacks, is offered. In addition, investigations highlighting the DL approach, cyber security applications, and dataset sources are described in this study. According to our findings, the restricted Boltzmann machine (RBM) achieves a classification accuracy of 99.72 percent when applied to a bespoke dataset, while the long short-term memory (LSTM) achieves a classification accuracy of 99.80 percent for the KDD Cup 99 data set. Furthermore the value of cyber security for dependable and practical IoT-driven healthcare systems is discussed in this essay.