Scientific Workflow
When conducting computational research, scientists do not normally think in terms of computers, databases, applications, and tools. Rather, they focus on the science in terms of physical concepts, theories, experiments, scientific data, and scientific analysis. When scientists become bogged down in the operational and technical details of the underlying computing environment, they are distracted from their true scientific pursuits. This represents a major barrier to the adoption of new computing infrastructure technology in the science laboratory.
An important premise in the concept of scientific workflow is that scientists need higher-level representations or abstractions to effectively employ scientific and computing resources in their research work. Thus, the goal of a scientific workflow system is to not only assist scientists in managing and executing their research processes, but also to translate those high-level scientific activities and concepts into low-level actions to be carried out on computing resources.
Rresearchers at PNNL are designing and developing scientific workflow systems that are capable of managing the use of distributed computers, instruments, applications, data, and other scientific and computation resources within the context of a scientific process.
PNNL researchers develop scientific workflow systems that
- Allow scientists to manage the execution of scientific processes
- Support ordering, scheduling, and execution of scientific and computational activities and steps within the encompassing scientific process
- Provide transparent access to underlying scientific and computational resources
- Provide a historical and reproducible record of process activities and steps
- Support both short-term and long-term research processes
- Allow scientists to collaborate on their scientific workflow representations and extend research by combining or merging them
- Facilitate the scientist's construction of scientific workflow representations and the linking of these representations to underlying resources
- Allow scientists and organizations to store, manage, and distribute workflow representations for different scientific research processes.
Contact: George Chin