A Hybrid Extreme Learning Machine And Symmetric DNA Cryptography For Preventing Information Leakage In Cloud Containers

Abstract

Cloud Container refers to a lightweight, portable, and executable software package and run applications in the cloud.  In cloud computing, containers are widely applied for deploying and managing applications since they offer portability, scalability, and consistency across different environments, better resource utilization than traditional virtual machines based data centers. Container cyber security is a primary concern as users and information grows in exponentially.  Sensitive information is highly vulnerable to cyber attacks during sharing or communication.  Several solutions are available to address the information leakage security in cloud container. However, confidentiality and data integrity remain the major concern. Therefore, a novel technique called cryptosystem called Hybridized EXtreme LEarning Machine based symmetric DNA cryptography (HEXLEM-DNA) is introduced.  The primary objective of the proposed HEXLEM-DNA technique is to strengthen the security of data storage in cloud container by ensuring a higher level of confidentiality and data integrity. Proposed HEXLEM-DNA technique operates through four key stages such as user registration, data encryption, authentication, and decryption. Initially, users submit their details to the cloud container for registration. Once registered, the cloud container generates a unique ID and symmetric cipher session key using Blum–Micali algorithm for each user. After that, the cloud data owner encrypts sensitive data before uploading it to the cloud container by using DNA-based encryption.  The encrypted data are uploaded into the cloud container for further processing. A cloud container acts as an isolated environment within the cloud infrastructure, helping to organize, manage, and protect the data. This approach significantly improves data confidentiality and adds a strong layer of security to cloud-based storage systems. When users access the data from cloud container, they must log in to the cloud container using these proper credentials for authentication. Authentication is performed using an extreme learning machine, which verifies whether the user is authorized or not. The cloud container grants the permission to the access the data for authorized user. Finally, the authorized user performs decryption with proper symmetric cipher session key using DNA-based decryption process.  This process facilitates secure data storage and communication with enhanced confidentiality and integrity in the cloud environment. The proposed method was evaluated based on various performance metrics, including data confidentiality, data integrity, and authentication accuracy, processing time, across different volumes of data and cloud users. The experimental results demonstrate that proposed outperforms traditional methods by achieving higher and confidentiality, integrity accuracy along with reduced processing time