XML-Based Computation for Scientific Workflows

Scientific workflows are increasingly used for rapid integration of existing algorithms to form larger and more comgrams. Such workflows promise to provide more abstract, yet executable views of data analyses through graphs of components representing computational tasks and edges representing the data-flow between tasks. However, designing workflows using purely data-flow-oriented models of computation introduces a number of challenges, including the need to introduce low-level components to mediate and transform data (so-called shims) and a large number of additional "wires" for routing data to components within a workflow. To address these problems, we present a modeling paradigm Virtual Assembly Lines (VAL) that can not only eliminate most shims and reduce the wiring complexity, but also provide significant help to scientists during workflow construction and maintenance. The core idea of VAL is to organize data on workflow channels into XML-like treestructures (i.e., simil...

• Type: Conference
• Year: 2010
• Where: ICDE
• Authors: Bertram Ludäscher, Daniel Zinn, Shawn Bowers
• Keywords: Additional Configuration Shell, Database, General Val Framework, ICDE 2010, Wiring Complexity