Pricing of Infrastructure as a Service in Cloud Computing

Please enable javascript in your browser.

Shahmoradi, Mohammad Hossein | 2025

0 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 57989 (01)
University: Sharif University of Technology
Department: Industrial Engineering
Advisor(s): Sedghi, Nafiseh
Abstract:
The growing demand for scalable and cost-effective computing resources has led to the widespread adoption of the Infrastructure as a Service (IaaS) model in cloud computing. One of the pricing models for resources, or virtual machines, in this service is the Spot model, which allows customers to access unused cloud server capacity at significantly lower costs compared to other models. However, despite their affordability, these virtual machines do not come with a Service Level Agreement (SLA) and may experience interruptions during use. Such interruptions increase the time required for customers to complete their tasks. In this study, we utilize discrete-event system simulations to examine the impact of enabling more accurate estimations of task completion times for these customers in the system. The results demonstrate that applying machine learning techniques to predict these completion times can provide significant benefits for both customers and service providers. Specifically, these predictions can enhance customer satisfaction, improve resource efficiency, and increase the revenue of service providers
Keywords:
Cloud Computing ; Infrastructure as a Service ; Discrete Event Simulation ; Machine Learning ; Spot Virtual Machines