Cloud computing is a network of remote computing resources hosted on the Internet that allow users to utilize cloud resources on demand. As such, it represents a paradigm shift in the way businesses and industries think about digital infrastructure. With the shift from IT resources being a capital expenditure to a managed service, companies must rethink how they approach utilizing and optimizing these resources in order to maximize productivity and minimize costs. With proper resource management, cloud resources can be instrumental in reducing computing expenses.
Cloud resources are perishable commodities; therefore, cloud service providers have developed strategies to maximize utilization of their resources. One method cloud providers employ is offering unused/excess computing resources at substantially discounted rates compared to other pricing tiers, whose pricing fluctuates with supply and demand levels. This is often referred to as spot pricing.
This study investigates methods to reduce risk and increase predictability of pricing for businesses utilizing Amazon Web Services (AWS) elastic compute cloud (EC2) Spot instance pricing tier by accurately predicting spot instance pricing over a specified time-frame using long short-term memory (LSTM) neural networks and comparing the results against traditional time-series Auto Regressive Integrated Moving Average (ARIMA) modeling. The results show LSTM model Spot Instance price predictions have an average reduction in mean absolute percent error (MAPE) of approximately 95 percent when compared to the baseline ARIMA model.
Lancon, Jeffrey; Kunwar, Yejur; Stroud, David; McGee, Monnie; and Slater, Robert
"AWS EC2 Instance Spot Price Forecasting Using LSTM Networks,"
SMU Data Science Review: Vol. 2
, Article 8.
Available at: https://scholar.smu.edu/datasciencereview/vol2/iss2/8
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License