How to Buy Cloud Resource Better for IaaS User: from the Perspective of Cloud Elasticity Testing

摘要

For IaaS users, how to buy cloud resources with the lowest spend is worth careful considering. Most IaaS providers utilize the elasticity feature in cloud computing to better provide resource allocation, however, IaaS users still know little about actual cloud resource allocation, so they have to apply for cloud resource very roughly to run their applications, which obviously costs more. In this paper, we address this issue from the perspective of cloud elasticity testing, that is, a more practical and better cloud resource purchase plan will be obtained in advance through elasticity testing of applications for IaaS users. We optimize the test generation towards existing elasticity testing by introducing genetic algorithm idea. Then evaluation metrics are designed to measure elasticity level of IaaS platform. After iterative and sufficient elasticity test executions, the best elasticity level for cloud applications is calculated from the test results to indicate better resource purchase plan. Test experiment results show that our method could help IaaS users to buy cloud resources more reasonable, which is much closer to the actual requirements of their applications and costs less.