Towards Efficient Scientific Data Management using Cloud Storage

Building more in-house datacenters to backup explosively growing scientific datasets is neither cost-effective nor in line with government green initiative. Cloud computing is emerging as a viable platform for data storage, collaboration and disaster recovery. We are going to develop a suite of "backup-to-cloud" tools that allows user to backup scientific datasets and applications into the cloud, and use cloud storage as a distribution platform. Our tool is optimized under technical and economical constraints posed by common cloud storage. We use both public and private cloud platforms to conduct feasibility study from performance, security, scalability and cost perspectives.