This project mainly aims to use cloud computing technologies for Internet of Things applications. Cloud, edge, and fog computing are the three levels that make up the Internet of Things network architecture. The fog layer is connected to over 25 million devices, increasing network data traffic. Consequently, there will be an increase in terrorist attacks. Therefore, it’s imperative to protect IoT data and recognize dangers. Because new attack types are always emerging, intrusion detection systems (IDSs) must be built to withstand these evolving threats. On the other hand, most contemporary intrusion detection systems (IDSs) are concentrated on predicting the probability of an assault, utilizing Deep Learning and Machine Learning principles. The two IDSs that are most often used are KNN and SVN. These intrusion detection systems have a high false alarm rate, a high time consumption, and a low accuracy in detecting attacks. Consequently, this paper presents a unique Deep Learning model that combines the Convolution Neural Network (CNN) and the Gated Recurrent Unit Network (GRU) methods. Moreover, the NSL-KDD and UNSW-NB15 datasets are employed. Multi-class classification is used to classify these datasets. The suggested model has a better detection rate than the other models. Compared to the other IDSs, its detection rate is 1.5 times greater. It also has a quicker detection rate and generates fewer false alarms. The results of the suggested model are ascribed to the application of machine learning and deep learning technologies.