The Global ATLAS Experiment: Advanced Research Networks Bring Scientists Closer to Solving the Biggest Mysteries in Physics
 Canadian researchers are taking part in a global physics project, recognized as the largest science experiment on Earth, which could uncover some of the great mysteries of the universe. What is dark matter? Why do objects have mass? ATLAS (A Toroidal LHC ApparatuS), the biggest, most complex particle detector ever built, could provide the answers to fundamental questions about the nature of matter and the forces that shape our world. The experiment, slated to become fully operational in late 2007 or early 2008, demonstrates the increasingly integral position that advanced global research networks and grid computing hold in leading-edge scientific studies.
| The Challenge: Analyzing Massive Amounts of Data on a Global Scale
The truly daunting task for researchers will be storing and analyzing the massive amounts of raw data produced by the experiment. The ATLAS data archive will take up 100 megabytes of disk space - every second. In one year of operation, an estimated 10 petabytes of data will be produced, equivalent to 200 million four-drawer filing cabinets full of text. Adding to the challenge posed by the volume of data is the fact that before scientists can even begin their analysis, these particle readings will have to be translated into information that is suitable for use in physics.
The unprecedented complexity and size of the ATLAS Experiment make it impossible for a single physics lab or team of scientists to manage the data. Roughly 2,000 researchers from 35 countries around the world have signed up to help the project move forward, giving ATLAS the truly global breadth implied by its name. |
|
About the ATLAS Experiment
ATLAS functions as an analysis tool for the massive particle accelerator called the Large Hadron Collider (LHC). The LHC, the world’s largest and most complex scientific instrument, produces and causes the collision of atomic and subatomic particles. The LHC represents the first time that particle collisions of such high energy have ever been possible, and only the ATLAS detector has the capacity to measure the particles’ properties.
Using ATLAS, scientists will monitor these collisions in hopes of confirming previously unquantifiable theories in particle physics. The experiment will also strive to make new discoveries, highlighted by the goal to recreate the conditions immediately following the Big Bang. This could shed new light on the theory that is used to explain the formation of the universe.
|
|
However, having the world’s biggest scientific team brings equally large management concerns. With data distributed at 150 facilities across six continents, the physicists, computer scientists and students involved with ATLAS will need to keep in constant collaboration to control access to the project.
The Solution: Advanced Networks and Grid Computing Enable Worldwide Collaboration and Data Analysis
Advanced research networks and grid computing are the fundamental building blocks that will allow ATLAS collaborators to conduct their work, from large-scale simulations to event reconstruction and analysis. Grid computing provides the infrastructure necessary for global data distribution and allows university and laboratory computers around the world to aggregate their CPU power.
These extremely high performance ‘grid’ computers are connected through regional, national and international advanced research networks, creating a large system managed by collaborating ATLAS sites. Because they enable the pooling of computational resources, grid computers deliver more power than most supercomputers and, no matter where researchers are located, they may log on to a grid and use the memory and processing abilities of all connected computers.
In BC, researchers connect to BCNET's Optical Regional Advanced Network (ORAN) and the national network, CANARIE’s CA*net 4. This in turn provides connections to international counterparts like Internet2 in the US. The capacity of these networks allows for the transmission of data almost instantaneously and provides direct point-to-point dedicated links between research facilities.
| "Canada is making a significant contribution to the computing resources of the ATLAS Experiment,” says Randy Sobie, Director of HEPNET/Canada, research scientist at the Canadian Institute of Particle Physics and adjunct professor at the University of Victoria (UVic). “The project would not be possible without the world-class networks provided by CANARIE and BCNET." |
This is largely related to the fact that TRIUMF, a Vancouver-based joint venture of several Canadian universities, will act as a primary Tier One analysis centre for the ATLAS Experiment. One of only 10 such sites, TRIUMF will store a large fraction of both processed and raw copies of data before they are distributed to Tier Two centres around the world for further analysis.
Particle physicist Mike Vetterli, a physics professor with Simon Fraser University (SFU) and a member of the TRIUMF ATLAS group, will lead the project to connect the Vancouver lab to facilities around the world. For this, BCNET and CANARIE have coordinated a dedicated link that transfers data between Vancouver, Toronto, New York, Amsterdam and the European Organization for Nuclear Research (CERN) in Geneva, Switzerland, where ATLAS is housed. The national Canadian ROADM network will provision TRIUMF with a dedicated 10 gigabit per second link to CERN and will also be used for transferring data to many Tier Two university sites, such as the universities of Toronto, Victoria and Alberta, as well as McGill and SFU.
The Results: Providing a Glimpse at the Future of Network-based Research
ATLAS construction is scheduled to be complete in June 2007. Later this summer, collaborators expect the project to commence data collection.
According to its participants, the significance of the ATLAS project reaches far beyond particle physics. Says Vetterli, “The grid used in the ATLAS experiment will do for large scale computing what the World Wide Web did for the global sharing of information.” Not only does it highlight the increasing emphasis that scientific research will put on teamwork and supercomputing, it reflects a future where network resources in distinct administrative domains will combine forces via advanced networks.
For more information visit http://www.gridx1.ca. |
