LDRD Seminar Series: ‘Data-intensive Computing and the Cosmic Frontier’
Ravi Madduri
Computer Scientist Ravi K. Madduri (MCS) and Cloud Services Team Lead Daniel Murphy-Olsen (CLS) will discuss their Laboratory-Directed Research and Development (LDRD) sponsored work at the LDRD Seminar Series presentation Tuesday, March 14, 2017.
“Data-intensive Computing and the Cosmic Frontier” will begin at 12:30 p.m. in the Bldg. 203 Auditorium. All are welcome to attend.
Abstract
We are exploring new computational technologies to deliver a dedicated data-intensive computational infrastructure that can be used for analyzing datasets from cosmological surveys and large-scale cosmological simulations. Our working hypothesis is that any practical implementation of a data-intensive system should consider managing data as a first-class design principle along with enhancements in advances in processor speeds and density. The basis of this capability is the Argonne-led Portal for Data Analysis services for Cosmological Simulations (PDACS).
Daniel Murphy-Olson
PDACS will be tuned to handle large-scale observational data streams and, very importantly, allow for joint analysis of observational and simulation datasets. PDACS will also provide a centralized index of collections of cosmology data across the DOE complex. Additionally, working with domain experts we will improve upon and extend analysis workflows available in PDACS. A key feature is the integration of software container technology, which will enable collaborative software development. This will improve software development, deployment process efficiency and ease integration of domain expert knowledge in high performance scientific software. As an important application, South Pole Telescope (SPT) workflows and analysis tasks will be integrated into PDACS. Finally, we will explore generalizing these tools so they can be leveraged in other science domains.
Biography
Ravi Madduri is a computer scientist in the Mathematics and Computer Science Division and Senior Fellow at the Computation Institute. He led several successful large projects in NSF, NIH and the U.S. Department of Energy. His research interests are in building sustainable, scalable services for science, reproducible research, large-scale data management and analysis. Madduri leads the Globus Genomics project, which is widely used for genomics, proteomics, and other biomedical computations on Amazon cloud and other platforms. He also architected the Globus Galaxies platform that underpins Globus Genomics and several other cloud-based gateways realizing the vision of Science as a Service for creating, maintaining sustainable services for science.
Daniel Murphy-Olson is the Cloud Services Team Lead in the Computing, Environment, & Life Sciences Directorate. He has over a decade of experience information technology operations and technical team leadership. He has spent the last four years also working as the lead system administrator for the KBase project. In this role he has become an expert on the deployment of cloud-enabled applications. He also has extensive expertise in the areas of storage and compute hardware system architectures.
