Combining Workflow Templates with a Shared Space-Based Execution Model

Abstract

The growth for scientific data has led to data analysis being a critical step in the scientific process. The next generation scientific data analysis environment needs to address two challenges i) productivity of the end-user and ii) scalability of the workflows. The need to ensure both goals requires us to revisit the design and implementation of workflow tools. In this paper, we study the interaction of Tigres and HOCL-TS towards meeting these goals. Tigres and HOCL-TS have evolved separately; however their complementary foci allows us to study these issues in greater detail. We describe the pros and cons of an approach that integrates Tigres and HOCL-TS and HOCL-TS extension to support common non-functional requirements such as logging and monitoring that can be made available to the users through the Tigres API.