AI-Based Cost Optimization and Security Compliance for Multi-Cloud Workloads

Singh, Aryan (2025) AI-Based Cost Optimization and Security Compliance for Multi-Cloud Workloads. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

Rapidly changing workload requirements and the emergence of complex cloud environments have led to the important issue of cost optimization in current cloud computing. The Autoscaling mechanisms Traditional mechanisms tend to use reactive threshold-based rules that tend to under-utilize the resources or delay the scaling actions. The proposed cost optimization path in this project is an intelligent, predictive, and policy-aware approach whose experiments utilize Long Short-Term Memory (LSTM) neural networks to predict relevant cloud workload metrics such as CPU utilization, wait time, and execution time. The model predicts the workload behavior in a time-series forecasting based manner, and initiates the autoscaling actions via an API interface built using SQLAlchemy, Flask, and Flask-Sqlalchemy. This system has a decision logic incorporated in it that automatically triggers the scale-up where the on-demand utilization level is expected to exceed 70 percent, or acts upon a scale-down in the case where the utilization level goes below 30 percent, and raises security and compliance alarms. The solution will emulate deployment environments native to the real world, that is, AWS Fargate and Azure App Service. Also, the LSTM predictive capabilities are compared to classic models SARIMA and Random Forest through MAE, RMSE and R 2 measure. Our work in this project provides a cloud-agnostic, scalable forecasting architecture, which integrates prediction precision, economic efficiency, and dynamically-capable enforcement policy, and therefore is applicable in next-generation cloud management systems.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Gupta, Punit UNSPECIFIED
Uncontrolled Keywords:	Cloud Cost Optimization; Time-Series Forecasting; LSTM; SARIMA; Random Forest; Autoscaling; Serverless Architecture; Multi-Cloud; Security Compliance; Flask API
Subjects:	Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	31 Mar 2026 09:37
Last Modified:	31 Mar 2026 09:37
URI:	https://norma.ncirl.ie/id/eprint/9269

Actions (login required)

View Item