Very Large Scale Data Management

Scientific Data Management at the Petascale and Beyond

Overhead view of the Large Hadron Collider at CERN (view as a high-resolution image)

Scientists involved in the Argonne Tandem Linear Accelerator System (ATLAS) experiment at the Large Hadron Collider at CERN—the European Organization for Nuclear Research—in Geneva, Switzerland, are seeking to understand the fundamental forces that have shaped the universe since the beginning of time and are examining those that will determine its fate. Questions concerning the origin of mass, extra dimensions of space, microscopic black holes, and evidence for dark matter candidates in the universe will be explored by colliding proton beams at near-Big Bang energies, using what is arguably the largest and most complex scientific instrument ever built.

The DIS Information Sciences Group, in collaboration with Argonne’s High-Energy Physics Division, has been integrally involved in the ATLAS computing and software project since its inception, taking the lead for defining and constructing a globally distributed, scalable, and efficiently navigable multi-petabyte physics data store.

Argonne leads an international team of developers that are providing the data store architecture and designing the navigational infrastructure, ATLAS control framework input/output infrastructure, and multi-terabyte metadata system. Argonne also provides support for distributed database servers and services. The data store and its supporting services operate in globally distributed grid- and cloud-based computing environments and at laptop scales, with performance and functionality tuned to specific high-energy physics access patterns and tool suites.

This global collaboration of more than 2,500 PhD-level physicists from 37 countries will generate several petabytes of raw data annually when collisions begin in 2010, with equivalent amounts of derived data and matching quantities of simulation data. The collaboration has already accumulated tens of petabytes of simulation and cosmic ray commissioning data in preparation for “turn-on” of the collider.

For more information, contact:
Craig Swietlik
Information Sciences Group
Decision and Information Sciences Division
Argonne National Laboratory
9700 South Cass Ave., Bldg. 900
Argonne, IL 60439
Phone: 630-252-8912
Fax: 630-252-5128
E-mail Craig Swietlik