Shwetha B, Manoj R, Hemanth V, Karthik S, Sai Charan, Vijay P

DOI: https://doi.org/10.5281/zenodo.23540123

Storing files, photos, bank records and office work on the internet (“the cloud”) is now a part of everyday life in India and across the world. This convenience also brings a danger: if a stranger gets into someone’s cloud account, that person’s private data can be stolen or misused. This paper explains, in simple words, a security design that protects cloud data in two ways. First, it checks who is logging in using several proofs together (password and PIN, fingerprint or face, a one-time code, and the way the person normally behaves). Second, it protects the data itself with locks that can change automatically when danger is sensed. A learning program (machine learning) watches every login, gives it a risk score, and decides whether to allow it, ask for more proof, or block it. The paper walks through each block of the architecture, explains how the blocks work together using a day to-day example, compares the design with the usual “password only” method, shows the screens of a working prototype called CloudShieldAI, and discusses limits, privacy and future work. The aim is that even a reader without a technical background can look at the architecture diagram and understand how it keeps cloud data safe

Cloud security, multi-factor authentication, adaptive cryptography, machine learning, anomaly detection, risk assessment, secure access.

The architecture explained in this paper protects cloud data from both sides. On the user side, it confirms identity using a password and PIN, a fingerprint or face, a one-time code and the user’s usual behaviour. On the data side, it uses locks that change with the danger, keys created from the user’s own proof, and constant watching for threats. A machine learning block connects the two sides by giving each login a risk score, and the intelligent response step turns this score into a clear action: allow, ask for more proof, or block. The strength of the design is that no single weakness is enough to open the data: a stolen password alone does not help the thief, and a changing lock does not help an attacker who has learned one method. At the same time, the honest user is troubled only when the situation really looks risky. The CloudShieldAI prototype shows how these ideas can appear in a simple website that ordinary people can use. With careful attention to privacy, speed and Indian legal requirements, this approach can make everyday cloud use safer for students, businesses and government services alike.

K. Sasikumar and S. Nagarajan, “Enhancing cloud security: A multi factor authentication and adaptive cryptography approach using machine learning techniques,” IEEE Open J. Comput. Soc., vol. 6, 2025, doi: 10.1109/OJCS.2025.3538557.
[2] P. Mell and T. Grance, “The NIST definition of cloud computing,” Natl. Inst. Standards Technol., Gaithersburg, MD, USA, Special Publication 800-145, 2011.
[3] P. A. Grassi et al., “Digital identity guidelines,” Natl. Inst. Standards Technol., Gaithersburg, MD, USA, Special Publication 800-63-3, 2017.
[4] S. Subashini and V. Kavitha, “A survey on security issues in service delivery models of cloud computing,” J. Netw. Comput. Appl., vol. 34, no. 1, pp. 1–11, 2011.
[5] V. Chandola, A. Banerjee, and V. Kumar, “Anomaly detection: A survey,” ACM Comput. Surveys, vol. 41, no. 3, pp. 1–58, 2009.
[6] A. K. Jain, A. Ross, and S. Prabhakar, “An introduction to biometric recognition,” IEEE Trans. Circuits Syst. Video Technol., vol. 14, no. 1, pp. 4–20, Jan. 2004.
[7] A. Ometov et al., “Multi-factor authentication: A survey,” Cryptography, vol. 2, no. 1, p. 1, 2018.
[8] J. Bonneau, C. Herley, P. C. van Oorschot, and F. Stajano, “The quest to replace passwords: A framework for comparative evaluation of web authentication schemes,” in Proc. IEEE Symp. Security Privacy, 2012, pp. 553–567.

Shwetha B, Manoj R, Hemanth V, Karthik S, Sai Charan, Vijay P (2026). Enhancing Cloud Security: A Multi-Factor Authentication and Adaptive Cryptography Approach Using Machine Learning Techniques. International Journal of Computer Techniques, 13(5), 554–561. ISSN: 2394-2231. DOI: https://doi.org/10.5281/zenodo.23540123