#SC09 Highlights - Monday Nov. 16th

Had a good day at SC09 in Portland today. I jumped between three different workshops that were going on today:

• WORKS09 : The 4th Workshop on Workflows in Support of Large-Scale Science
• MTAGS09 : The 2nd Workshop on Many-Task Computing on Grids and Supercomputers
• Using Clouds for Parallel Computations in Systems Biology

Overall the presentations were pretty good. Each of the workshops had at least one paper centering around the use of Hadoop/MapReduce. Below are the list of papers/presentations that I found to be most interesting:

• "Making Infrastructure Invisible" - Dr. Dan Reed, Microsoft Corporate VP for Extreme Computing (MTAGS09 Keynote)
  • "Successful technologies are invisible."
  • "High value does not imply high utilization. Rapid response is often more important."
• "Robust Workflows for Science & Engineering" - David Abramson et. al (MTAGS09 Invited Paper)
  • Talked about Nimrod/K, the workflow engine for the Nimrod Toolkit. It is built on Kepler.
• "Lessons Learned From a Year's Worth of Benchmarks of Large Data Clouds" - Yunhong Gu and Bob Grossman (MTAGS09 Invited Paper)
  • Talked about Sector/Sphere - An open source high performance distributed file system and parallel data processing engine. Performance was shown to be significantly better than Hadoop.
  • Demos of Sector/Sphere are being given at the National Center for Data Mining (NCDM) Booth (#1309)
    
• "Cloud Computing for Science" - Kate Keahey (Systems Biology Workshop)
  • Talked about Nimbus and Science Clouds

In the evening I attended the Opening Gala. Just took a browse around the exhibition floor tonight, so don't have any booth specific comments to add at this time. Will comment on booths/projects I found interesting in future blog posts this week.

Something else that is pretty cool and worth checking out is the Need4Feed site hosted by Purdue University. It is following the #SC09 hashtag on twitter and generating various rankings and statistics.

Tomorrow I will be giving a demonstration of GeoChronos - an on-line collaborative platform for earth observation scientists leveraging social networking, cloud computing and Semantic Web technologies. Come and check it out between 2pm - 3pm at the Compute Canada - CANARIE Booth (#142).