Workshop on In Situ Data Management

Sponsored by the U.S. Department of Energy, Office of Advanced Scientific Computing Research (ASCR)
North Bethesda, MD
January 28 – 29, 2019

Program Manager: Laura Biven

ISDM Brochure

Download the ISDM brochure at the Advanced Scientific Computing Research (ASCR) Community Resources page.

The U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research (ASCR) will host the Workshop on In Situ Data Management at the Bethesda North Marriott.  The workshop will begin at 8:30 AM on Monday, January 28 and will end approximately at 3:00 PM on Tuesday, January 29, 2019.

This workshop seeks community input on the development of in situ capabilities for managing the execution and data flow among a wide variety of coordinated tasks for scientific computing. The goal of this workshop is to focus the community on the development of in situ data management in service of the DOE mission, by identifying future opportunities and capabilities needed for in situ data management and the associated research challenges in meeting these.

In situ data management (ISDM) consists of the practices, capabilities, and procedures to control the organization of data and enable the coordination of and communication among heterogeneous applications, executing simultaneously in an HPC system, cooperating toward a common objective. The workshop considers ISDM beyond the traditional roles of accelerating simulation I/O and visualizing simulation results, to more broadly support future scientific computing needs. In particular, the convergence of simulation, data analysis, and artificial intelligence will require machine learning, data manipulation, creation of data products, assimilation of experimental and observational data, analysis across ensemble members, and, eventually the incorporation of tasks on non-von Neumann architecture. Plenary and breakout sessions at the workshop will dive into ISDM topics such as science applications, computing systems, data and communication models, programming and execution models, provenance and reproducibility, analysis algorithms, and software productivity.

The DOE Office of Science supports some the world's largest experiments, observations, and simulations in order to answer fundamental science and engineering questions across the DOE mission. The data life cycle of these grand-challenge campaigns can be formidable to navigate, communicate, and reproduce. In the future, scientific computing will increasingly incorporate an even greater number of different tasks that need to be managed. Improving the automation, usability, reproducibility, and efficiency of data management in this rich space of applications and architectures can increase scientific productivity and validate conclusions from scientific data.  ISDM is therefore a critical capability for enabling the complex scientific computing workflows needed to address DOE mission needs, and your active participation in the workshop discussions will help shape future ASCR research directions in this area.

Organizing Committee

Tom Peterka, Chair (ANL)
Debbie Bard (NERSC/LBNL)
Janine Bennett (SNL)
Wes Bethel (LBNL)
Ron Oldfield (SNL)
Line Pouchard (BNL)
Christine Sweeney (LANL)
Matthew Wolf (ORNL)