Research Computing

Dartmouth Joins Aristotle Cloud Federation

Building a model of cloud computing resource-sharing in support of multi-discipline scientific workflows

• By David Raths
• 11/08/18

Dartmouth University has joined the Aristotle Cloud Federation, the federated cloud-based research computing project established by Cornell University and its partners, the University at Buffalo and the University of California, Santa Barbara. Federated cloud computing models promise to make it easier for multi-institutional academic research projects to aggregate, share and analyze large data sets.

The three founders of Aristotle are exploring the computing model under a National Science Foundation Office of Advanced Cyberinfrastructure award. They hope it will serve as a model that other institutions can follow in terms of providing flexibility in support of multi-discipline scientific workflows, elasticity by sharing resources, and an allocation model that provides a fair exchange mechanism for resource access between and across multiple institutions.

"We're excited about this opportunity," said Mitchel Davis, vice president and chief information officer at Dartmouth, in a prepared statement. "A federated cloud model has the potential to facilitate resource-sharing between campuses and is emerging as an important consideration in cyberinfrastructure ecosystem planning."

Davis said he hopes to provide Dartmouth researchers with access to cost-effective, on-demand systems, software and data in a hybrid fashion that will let them choose from private, public and federated cloud offerings.

Dartmouth is the first university outside the Aristotle developer sites to join. Other NSF-funded testbeds for federated cloud computing include Chameleon, a collaboration of the University of Chicago and UT-Austin's Texas Advanced Computing Center; and Cloudlab, a collaboration of the University of Utah, Clemson University, and the University of Wisconsin.

Aristotle researchers are developing a new tool for cloud metrics to make online forecasts of future performance and allocations levels available to Aristotle users. This tool will display job accounting metrics over any desired timeframe, as well as trend analysis results, to predict when it is appropriate to burst from local resources.

Use cases span a variety of fields, including earth and atmospheric sciences, finance, chemistry, astronomy, civil engineering and genomics. Collaborators from more than 40 academic institutions, public agencies and research labs, as well as citizen scientists, are participating. For instance, one project involves the integration, visualization and analysis of geospatial data from around the world.

David Lifka, vice president for information technologies and chief information officer at Cornell and Aristotle's principal investigator, said the federation is looking forward to working with Dartmouth to bring science and digital humanities into the project. In a statement, he added, "New use cases provide us the feedback we need to harden federation technologies and fine-tune our documentation."