Abstract:

Currently more and more enterprises transfer their core business and data to the cloud, and a number of the business needs elasticity service to deal with the real-time changing of workload. So there is an urgent need to evaluate elasticity, however, we lack a relatively overall evaluation method. To solve this problem, we conduct a comprehensive analysis from the perspectives of resource allocation, QoS and resource allocation time, and propose an evaluation method that can be applied to both providers and users. Moreover, we design computation models for resource allocation and resource allocation time to improve the existing penalty model. Finally, TPC-W benchmarks are tested on the CloudStack platform with two elasticity scaling strategies, namely auto-scaling and scale-out, which verifies the proposed method.

Key words: elasticity service, evaluation method, resource allocation, QoS, resource allocation time