A Novel MWKF-LSTM Based Intrusion Detection System for the IoT-Cloud Platform with Efficient User Authentication and Data Encryption Models

Abstract During the past decade, the Internet of Things (IoT) integrated Cloud (IoT-Cloud) has gotten a lot of attention. Security turns out to be an important issue in the IoT-Cloud scenario as millions of devices is connected via the internet. Developing an intrusion detection system (IDS) along with a prevention model to improve IoT-Cloud security, with considered delay, detection rate, as well as false-positive rate are the most important research problems. Here, a new deep learning (DL) model, Morlet Wavelet Kernel Function included Long Short Term Memory (MWKF-LSTM) classifier is proposed to recognize the intrusions in the IoT-Cloud environment. Initially, to maintain a user’s privacy in the network, the SHA-512 hashing mechanism incorporated blockchain authentication (SHABA) model is developed that checks the authenticity of every device/user in the network for data uploading in the cloud. After successful authentication, the data is transmitted to the cloud through various gateways. Then the IDS using MWKF-LSTM is implemented for identifying the type of intrusions present in the received IoT data. The MWKF-LSTM classifier comes up with the Differential Evaluation based Dragonfly Algorithm (DEDFA) optimal feature selection (FS) model for increasing the performance of the classification. After intrusion detection (ID), the non-attacked data is encrypted as well as stored in the cloud securely utilizing Enhanced Elliptical Curve Cryptography (E2CC) mechanism. Finally, in the data retrieval phase, the authentication of the user is again checked for ensuring user privacy and preventing the encrypted data in the cloud from intruders. To show the efficacy of the proposed research model, simulations are performed and the outcomes prove the superior performance of the presented approach over existing models.