BNL’S Computational Science Initiative: A Co-design Approach to Supporting Data Intensive Science at Large-Scale Experimental Facilities
Kerstin Kleese-Van Dam, Brookhaven National Laboratory
Tue 10:45 – 11:30am
BNL provides one of the largest data analysis and storage capabilities for scientific data worldwide. With currently just over 90PB of archived data we are the 2nd largest archive in the US (following NOAA with 100PB) and the 4th largest in the world. We currently ingest 30PB/months, distribute 32PB/months and process 150PB/months of data. This data largely stems from high energy and nuclear physics experiments and their accompanying analysis and interpretation. With the opening of the newest US synchrotron – NSLS II – at BNL in 2015, we are expecting to significantly increase our data capabilities. The NSLS II is the brightest synchrotron worldwide and will comprise of 60 different end stations in its full build out. The instruments deployed at the facility utilize the latest detector technologies and create orders of magnitude more data than any other facility of this type. To enable scientists to fully leverage the scientific capabilities of this facility we need to provide them with novel, high performance streaming data analysis and large scale data analysis and interpretation capabilities. This talk will present the BNL Computational Science Initiative and its co-design approach to researching and creating the required hardware, methods and tools necessary to meet this challenge.