JUNE 18–22, 2017

Session Details

Name: Big Data Experiments & Big Data Analysis (Driving the Convergence of Big Compute & Big Data in HPC)
Time: Wednesday, June 21, 2017
11:00 am - 12:30 pm
Room:   Panorama 2
Messe Frankfurt
Breaks:12:30 pm - 01:45 pm Lunch
Chair:   John Shalf, LBNL
Abstract:   The dramatic rise of “Big Data” and Machine Learning in commercial data centers has driven a revolution in data-driven discovery in the business analytics community. The common need for large scale parallel computing in both HPC and for “Big Data” has led to a vision of Data-HPC convergence in the market place. However, the killer application to drive such convergence has remained elusive, but the killer application may well be experimental sciences, which historically not had as large of a presence in HPC centers as theoretical science (e.g. modeling and simulation).
Numerous Big Data Experiments require Large Scale Simulations on HPC Systems just to handle their growing data management and analysis challenges. Much of this is driven by the double-exponential improvements in sensor technologies for sensors and large scale experimental apparatus (improvements in CCD resolution and data acquisition rates, and explosive pace of improvement in DNA sequencers), and the rise of science that mines data from pervasive IoT sensor networks. The rise of big data in experimental science is transforming the service model for HPC centers and is having a substantial impact on how HPC systems are architected.
The purpose of the session is to explore how HPC centers are re-inventing themselves to bring together both experimental data and theoretical models. We seek to demonstrate the consequences/challenges of the Big Data tsunami resulting from large scale experimental science, and how it affects the architecture and even the fundamental service model of future HPC centers. Our three distinguished panelists will cover different perspectives of applying big compute to handle big experimental data for discovery science.
Presentations: A Superfacility Model for Data Intensive Science
11:00 am - 11:20 am
  Kathy Yelick, LBNL & UC Berkeley
"Think Big - Think Outside the Box!" - Extreme Computing Requirements for the Square Kilometer Array (SKA)
11:20 am - 11:40 am
  Ronald P. Luijten, IBM Research Zurich
ABCI - AI Bridging Cloud Infrastructure
11:40 am - 12:00 pm
  Satoshi Matsuoka, Tokyo Institute of Technology
Questions & Answers
12:00 pm - 12:30 pm