Abstract

A national research Cloud is being created as part of the National eResearch Collaboration Tools and Resources (NeCTAR) project, to supply a cost-effective computational infrastructure to Australian scholars. Since it is hard to comprehensively understand the available Cloud resources through the open indicators, we decided to empirically investigate the infrastructural performance of the NeCTAR Research Cloud. In particular, our current evaluation work shows that the communication data throughput vary significantly within and between the five existing Cloud nodes. Although operated as one Cloud system, the NeCTAR Research Cloud might have employed heterogeneous network equipment and/or technologies. As the NeCTAR project is still ongoing, the one-off evaluation results could be quickly out of date, and the evaluation implementations should be frequently repeated or replicated for monitoring or consuming the Cloud services. Thus, unlike the existing evaluation studies that mainly focused on the evaluation results or tools, we emphasize the complete procedure and backend logic of evaluation implementations. Practically, as demonstrated in this paper, the procedure and logic can be recorded into Cloud Evaluation Experiment Methodology-based evaluation templates to facilitate repeatable and comparable Cloud services evaluation.