CINET

Working with collaborators at Virginia Tech and Indiana University, we will explore automated mechanisms for configuring virtual machine (VM) images to run on cloud research at Argonne National Laboratory. The VMs will be designed to allow both storage of integrated data sets, for use by large open communities, as well as on-demand computational resources. The integration of this on-demand resource is key to easy deployment and use, and we will explore APIs and remote interfaces that allow CINet users to access, start, and stop resources as the work. Finally, while many VM environments use single node images for cloud computing, we expect that CINet users will need larger computational resources, and will explore the automatic integration of MPI-based multi-node VM environments by utilizing the Nimbus infrastructure.