Dynamic Duplication Attack Detection and Mitigation System (DDADMS)

Objectives: To propose a suitable approach to detect, analyse and prevent duplication attacks in cloud computing for effective and secure data management. DDADMS strengthens cloud security and dynamically manages resource usage with real-time de-duplication, machine learning profiling and LSTM prediction. Methods: This study has proposed Dynamic Duplication Attack Detection and Mitigation System (DDADMS) which uses machine learning, hashing, and pattern matching to detect duplicate data when file names and also to change metadata. It develops normal duplication profiles from end-to-end software analysis on client hosts and cloud hosts. Real-time de-duplication is combined with LSTM networks for duplication attack prediction using past patterns for proactive defence mechanisms against advanced threats. Findings: DDADMS was compared against three others Signature-Based Detection and Behavioural Analysis System, Traditional Hashing Technique, and Behavioural Analysis System and outshined them far better. Traditional Hashing Technique can detect duplication attacks 80% of the time but at a comparatively higher false positive rate of 10.5%, water down its differentiation between malicious and benign duplication. Signature-Based Detection increases detection to 85% but at a loss of 8.0% false positive. Behavioural Analysis System identifies 88.5% at a false positive rate of 6.5%, higher than the first two but lower than DDADMS. A mean de-duplication time winner for the Behavioural Analysis System, DDADMS has a higher performance in all three methods in overall detection accuracy, optimality of resources used, and threat blocking. By projecting LSTM-trained models onto huge history datasets, DDADMS forcefully avoids and counteracts duplicate attacks, curbing wastage of storage and redundancy of data by a huge extent. Thus, DDADMS strengthens cloud security and dynamically manages resource usage with real-time de-duplication, machine learning profiling and LSTM prediction. Novelty: DDADMS brings forth a revolutionary innovation in cloud security with real-time duplicate elimination and predictive analysis integration. With other available mechanisms reacting to duplication attacks only after their occurrence, DDADMS identifies duplicate data the instant it is present and deletes it, achieving maximum storage optimization with superior detection rates. The innovative approach provides adaptive and dynamic defence strategy against dynamic duplication attacks, making DDADMS a very powerful cloud security management solution. Keywords: Machine Learning, Attack Detection and Mitigation, Cloud Security, Real-Time De-Duplication, Machine Learning Profiling, LSTM Prediction, Resource Optimization