September 2, 2010

Hybrid Workflow Model at Heart of Tornado Research

Nicole Hemsoth

Researchers examining how tornadoes form have turned into “twister chasers” rushing from site to site to get their equipment on the ground before the tornado hits. Their efforts are aimed at more accurately predicting when touchdown will occur to improve evacuation and warning, but this is, as one can imagine, quite a dangerous and troublesome process. Simply getting to the cell’s site at the right time is a challenge for the VORTEX2 group, thus other compute-based esearchers are looking to distributed computing to take on the challenge and reduce the complex of fieldwork.

The Center for Analysis and Prediction of Storms undertook a hybrid computing project called the “Linked Environments for Atmospheric Discovery II” (LEAD II) with the help of the Big Red cluster and Microsoft’s Azure to better predict tornado conditions and to aid in the advancement of VORTEX2 goals.

Although LEAD II is only a fraction of the larger VORTEX2 project, part of what makes it remarkable is that it signals the first use of a hybrid workflow model for the participants, which was created using Microsoft’s Trident Scientific Workflow Workbench to handle the front-end workflow system, which then doled out pieces of it to backend Unix and Linux-based resources, including Indiana University’s Big Red.

As LEAD II researcher Beth Plale explained, “A lot of scientists use Windows tools such as Excel…We think that utilizing a Windows workflow system on a Windows box is a step towards providing broader flexibility, because of this affinity of a lot of scientists to use Excel and because of the emergence of the cloud-based Azure platform.”

Overall, the team decided that the use of the hybrid workflow model for the project was a success due to the added flexibility it offered, not to mention a great exercise in experimenting with different models for use in future projects.

Full story at International Science Grid This Week

Share This