The study of multi-instance purchase decision-making for minimising customers' cost under fluctuating cloud demands. (2016)

• Record Type: Journal Article
• Authors: She, Qiping; Li, Qing; Deng, Juan
• Is Part Of: International journal of high performance computing and networking. Volume 9:Number 1/2(2016)
• Page Start: 127
• Page End: 133
• Publication Date: 2016
• Subjects: cloud computing -- demand fluctuation -- purchase decision making -- multiple instances -- Amazon EC2 -- risk losses -- customer costs; High performance computing -- Periodicals; Computer networks -- Periodicals; High performance computing; Periodicals; 004.05
• Languages: English
• ISSNs: 1740-0562

Abstract

At present, cloud providers can offer cloud consumers three provisioning plans for computing resources, namely on-demand instance, reserved instance and spot instance. On-demand instance is stable, but very expensive; reserved instance is also guaranteed, and advance reservations are required, however customers' future demands are often uncertain; price of spot instance could be the cheapest, but customers have to face the risks of instance interruption. Therefore, how to minimise customers' total cost with taking into account the purchase costs and risks is one of the problems needing urgently to be solved. Taking Amazon EC2 as an example, this article makes a study of multi-instance purchase decision-making for minimising customers' cost under fluctuating cloud demands with taking risk losses into consideration, then describes the conditions and the reasons for optimal decision with a case in detail.