Cloud Computing

Regional Project Targets Data Storage for Research Cloud

• By Dian Schaffhauser
• 09/24/15

The on-going challenge of figuring out how to store and share the data generated in research projects will receive $4.9 million worth of attention in a project undertaken at four institutions. The University of Michigan, Michigan State U, Wayne State U and Indiana U are collaborating on a new kind of petabyte-scale data system.

Multi-Institutional Open Storage Research InfraStructure (MI-OSiRIS) will combine object-based software-defined storage technology with a network infrastructure to give users a distributed storage system that they can access from any of their institutions. It'll run on commodity hardware.

The data system will run Ceph, a massively scalable open source storage platform that supports multiple storage interface approaches and allows for configuration of access, replication, distribution and integrity on a domain basis. The software-defined storage aspects of Ceph will enable the system to self-manage operations, such as pinpointing the "optimal network path" between a scientist and the data, wherever it may be stored.

The infrastructure will use each school's existing authentication to let researchers give controlled access to their data across universities. Along the way, the details of design, code and operational experiences will be documented and published to allow MI-OSiRIS to be replicated elsewhere by other groups of cross-institutional collaborators.

"What we're trying to do here is expedite the time to discovery," said Shawn McKee, a research scientist in physics at U Michigan's College of Literature, Science and the Arts. "Scientists should be able to focus on their science without having to become experts in data management."

McKee is one of numerous researchers who will test the system in projects exploring ocean modeling, biostatistics, cancer, degenerative diseases and aquatic biology. Another is Hiroko Dodge, professor of neurology at the U Michigan Medical School, and colleagues at Wayne State, who will employ MI-OSiRIS in their study of Alzheimer's. Sensors in the homes of seniors gather 24/7 information about their walking speed, sleep patterns and computer and phone usage. That data is combined with additional data from cognitive tests, MRIs and other areas.

"Direct access to data between our sister institutions will eliminate hours and even days lost copying massive files from one place to another," said Patrick Gossman, deputy CIO for research at Wayne State. "The end result will be improved research productivity in health, aging, the environment and other areas important to us all."

The new infrastructure for MI-OSiRIS will be built by U Michigan's Advanced Research Computing-Technology Services. The work is being funded by the National Science Foundation.