CI-Miner: Scientific Workflows in Support of Cyber-Infrastructure Functionalities

Initial cyberinfrastructure (CI) efforts have primarily focused on gathering and recording scientific information from distributed, cross-disciplinary domains and developing customized applications to access such information. CI-Miner is a framework for enhancing an existing CI with knowledge that can be used to develop scientific applications that knows how to use the resources available over the CI. Instead of focusing on a set of functionalities implemented statically to access CI resources, as often done through web-accessible portals, CI-Miner provides a framework that allows the scientist to use functionalities that may be dynamically realized from a scientist’s requirements. In this case, both application capabilities and CI resources used by applications are specified in terms of concepts defined in their specific scientific discipline domains. The types of applications supported by CI-Miner are scientific workflow applications. CI-Miner provides a framework, tools and an initial method to allow scientists to define workflows (i.e., an ordering of a collection of activities) that realize a scientific task of interest. By empowering the scientist to define workflows to accomplish their scientific endeavors, CIMiner improves the utilization of CI resources towards the advancement of science by allowing scientists to experiment by building rapid prototypes, discovering alternate resources to accomplish a task, and transferring knowledge faster by grounding current scientific endeavors on other well established resources. Furthermore, empowering scientists with the capability of defining their own applications of interest is beneficial for CI resource development efforts because it allows the scientific efforts to drive resource development by identifying CI resources required to realize the specified applications, but which may not be currently implemented. The CI-Miner framework is based on a collection of technologies where some of the technologies have been developed by our own group while others have been developed by other groups of the Semantic Web and Grid Computing communities. By capturing domain knowledge from the scientist, CI-Miner can also provide scientists with guidance and examples on how to use CI-Miner technologies to build applications that rely on CI resources.