Cryptography and Adversarial Analysis for Cloud Data Security

Cloud computing, a novel paradigm for distributed data processing, has emerged as a transformative technology enabling the storage and retrieval of data over the Internet. The prevalence of Software-as-a-Service has revolutionized work dynamics, fostering a culture of secure remote access across diverse devices. Within the realm of Cloud Storage, the imperative for effective security and privacy measures is underscored. This research proposes a robust data security framework employing the Cryptography Splitting Algorithm. User files undergo encryption, increasing security levels from 25 to 75 are securely stored with a cloud service provider. The Cloud Manager empowers users by enhancing security, managing tracking mechanisms, and safeguarding metadata, thus fortifying data confidentiality and flexibility in utilization. Leveraging the Cryptography Splitting technique ensures secure data storage, concurrently enhancing both security and privacy for user data. Furthermore, the proposed methodology undergoes rigorous scrutiny in an adversary setup, incorporating neural network analysis to evaluate the resilience of data security measures. This analysis reveals the encryption scheme protects communication by maintaining a low $A-B$ loss around 0.1 while reducing Eve’s ability to intercept messages from 0.7 to 0.2. This research contributes a comprehensive approach to securely managing data in the dynamic landscape of cloud computing.