FEMCRA: Fine-Grained Elasticity Measurement for Cloud Resources Allocation

摘要

It is indispensable for a cloud platform to provide flexible elasticity service. However, the cloud users do not know whether elastic resource allocation of cloud platform matches their resource requirement or not, and the inappropriate purchase plan will have an impact on effect of elasticity service. Thus, a fine-grained and suitable purchase plan of cloud resources is inevitably needed. In this paper, we propose a fine-grained elasticity measurement method for cloud resources allocation towards upper-level cloud applications, called FEMCRA. It is a feasible measurement method from the perspective of testing cloud applications in advance to find a fine-grained and suitable elastic resource allocation scheme. We construct an integrated environment for simulating various elasticity level scenarios based on OpenStack, and data analysis application from CloudSuite is deployed to act as applications under testing. By executing that application under different elasticity rule-sets, which is a group of testing strategies that makes the cloud platform to be automatically scaled, we get the optimal elasticity level with least quantity of cloud resource, and the best purchase plan is correspondingly obtained which could save the expense for cloud users.