As cloud services are deployed, users will have a greater selection of choices. Since cloud services are charged based on usage rather than purchase, service migration can occur any time if the current service does not provide an acceptable quality of service (QoS). This means that cloud service providers need to manage their offered QoS to keep their customers satisfied.

A main contribution of this paper is its presentation of a novel user-side cloud service management framework that includes service discovery and QoS monitoring, QoS forecasting and early warning, and decision support by recommending a service with the best QoS. The framework registers all available services in the cloud environment as potential candidates for cloud service users. The authors applied the proposed framework to five Amazon Elastic Compute Cloud (EC2) infrastructure as a service (IaaS) cloud services.

The QoS criteria are not studied well in this paper. Only typical criteria such as central processing unit (CPU) response time, memory response time, and input/output (I/O) response time are considered as examples. Although a user feedback mechanism is mentioned in the service monitoring module, the paper does not present how the feedback will be incorporated into the calculation of QoS deviations.

This paper spends quite a lot of pages on explaining existing forecasting methods, while the proposed fuzzy inference system is briefly mentioned. Thus, readers can wonder how the fuzzy membership functions are determined. Although a de-fuzzification result is presented, it is difficult to understand how the de-fuzzification is performed due to lack of explanation. Furthermore, it is difficult to agree on whether the fuzzy mechanism is acceptable, since the paper does not show any comparison with other mechanisms.