Computing Systems Seminar

Transcript

1. Designing research-computing systems that respond to the varying needs of

   ILRI researchers Alan Orth ILRI, Kenya December 5, 2013

2. Research-computing platform Must be research agnostic and flexible to cater

   to ILRI’s diverse disciplines

3. Diverse disciplines Genetics Statistics Geoinformatics Molecular Biology Immunology Breeding Ag

   Economics Value Chains Bioinformatics Social Science Epidemiology Computer Science

4. Varying needs Sequence data Genome assembly Geospatial data Databases ILRI

   films Phenotypic data System monitoring Mobile data collection Metagenomics pipelines Web servers

5. Wide scope Cows Zoonosis National partners Sheep Women Markets Baseline

   surveys Increasing productivity Value chains Cassava

6. Support ILRI scientists and collaborators wherever they may be

7. Primary characteristics Computational capacity should be high and fast Data

   should be accessible and safe

8. Provide a strong foundation for... Cutting-edge science supported by a

   state- of-the-art infrastructure. Building a community of practice by teaching good habits and helping scientists use the tools. Breaking down silos and promoting collaboration via openness and sharing.

9. Standing on shoulders of giants

10. Embrace “open” Open-source software Open-access science

11. Building “next-gen” research computing Four years at ILRI ✓ Three

    iterations of HPC ✓ Create homogenous cluster environment ✓

12. Alan at ILRI (2010 - 2013)