Supporting Scientific Workflows Through First-Class Connectors

Student: David Woollard, USC

Advisor: Nenad Medvidovic, USC/ISR

Abstract: Scientific workflows (models of tasks, data and dependencies) are process models that have greatly aided scientists in managing scientific experimentation via computer simulation and data analysis– so-called “in silico” processes. Production workflows are in silico processes that use scientific algorithms or processes to produce large amounts of data, emphasizing robustness, high throughput, and automation of tasks. While production workflow systems require both scientific understanding and significant engineering to meet their requirements, there is no current methodology for separating out production functionality from scientific software in such a way as to allow the scientist to manage overall scientific behavior whilst the software engineer manages the production aspects of the system. A software architecture–the elements, form and rationale of the system design–allows scientists and engineers to communicate at a level of abstraction that provides each expert with the necessary view of the workflow task without requiring each to become an expert in the other's domain.

Bio:

David Woollard is a PhD candidate at the University of Southern California studying the role of software architectures in high performance scientific code as part of the Software Architectures Group led by Dr. Neno Medvidovic. Additionally, David is a Software Engineer at NASA's Jet Propulsion Laboratory. At JPL, David has served in the roles of developer, architect, and system engineer on the Orbiting Carbon Observatory and NPP Sounder Peate missions.