Towards universal cloud service for distributed large scale scientific data

Integrating heterogeneous scientific dataset and providing universal collaborative services is an important goal for e-Science project of Chinese Academy of Sciences (CAS). In this paper, we present a universal cloud service for scientific dataset applications. We firstly design a resources pool container to support distributed heterogeneous scientific resources integration and continuous expansion. We then provide a scalable architecture to support application virtualization, which aims to support cloud service invocation applying back-end transparent resources pool. Finally an open interface is presented for third-party to develop customized applications. We give the experiments on the cloud service infrastructure of its performance, which shows our approach achieves considerable efficiency.