Efficient Intrusion Detection and Prevention Model in Cloud Environment Using Sgd-LSTM and C2HA

Abstract

: Cloud computing is an attractive technology paradigm that has been widely used as a tool for storing and analyzing the data of different users. Since access to the cloud is achieved through the Internet, data stored in clouds is susceptible to attacks from external as well as internal intruders. Henceforth, cloud service providers (CSPs) need to take action in order to provide a secure framework that would detect intrusion in the cloud and protect and secure customer information against hackers and intruders. This paper proposes a Sgd-LSTM and signature-based access control policy based Intrusion Detection and Prevention System (IDPS) model which is meant to detect and prevent various intrusions in the cloud. The proposed system includes three phases: the user registration phase, intrusion detection phase, and intrusion prevention phase. Initially, user registration is performed based on a unique ID and password, and then, the password is converted into hashcode by using the C2HA algorithm and then stored in the cloud for authentication purposes. In the intrusion detection phase, the status of cloud data is predicted by employing the Sgd-LSTM classifier in order to discard the intruder data packets from the cloud. At last, in the intrusion prevention phase, data access to the cloud environment is controlled by using signature-based user authentication in order to authenticate the legitimate user. The proposed classifier can effectively detect the intruders, which was experimentally proved by comparing it with the existing classifiers.