Opinion - Me, my friends, our grid: bringing people together for great science
Frank Würthwein is a particle physicist with University of California San Diego and a user of Open Science Grid. He is also the OSG applications coordinator and a member of the CMS collaboration at the Large Hadron Collider, and says grids are as much about sociology as they are about science.
As the clock ticks towards startup for the Large Hadron Collider, particle physics draws ever closer to what many hope will be a revolution for the field: a giant step forward in our understanding of the universe.
For many physicists, the greatest fear is that we will find the Higgs Boson, and that in the wake of this success, nothing will change.
More exciting, more fascinating, is the possibility that finding the Higgs will bring with it a new paradigm in physics, propelling us forward into an entirely new direction of research.
And so, in a bid to power the LHC’s computing requirements, we are taking leading edge technology, bleeding edge technology, and applying it in more than a hundred sites around the globe.
If we had tried to build grids ten years ago it would’ve been impossible. If we were to try in ten years time, it wouldn’t be a problem. It is because grid technology is evolving right now—in a race to keep pace with the needs of high energy physics—that it poses such great challenges, and represents such a tremendous achievement.
Coming to the table
Collaborations such as Open Science Grid and EGEE have put this grid technology on the table. Now it’s relatively easy for everybody on the planet to say, “Hey, let’s bring this science home. We no longer have to send our scientists halfway round the world to access this stuff.”
Some of these sites already have existing infrastructure and just want to interface; others are starting completely from scratch. In some cases we’re working with countries that lack infrastructure, resources and expertise. Many of the IT departments we’re working with have never done anything at this scale or complexity ever before.
And yet, these sites are not motivated by IT; they are motivated by science.
IT departments are now approaching grid collaborations, but they’re not doing it for the Large Hadron Collider; they’re not doing it just for physics. They’re coming because joining a grid is a strategy: it’s a way to grow, to more efficiently satisfy their own communities, and to bring new communities to their cluster.
Open all hours
This breadth of users and resources means that collaborations like Open Science Grid need to stay flexible if they are to remain truly “open” and inclusive.
The biggest challenges in grid technology are not technological; they’re social. Building new communities, where people learn how to share and work together, is as important as building new computer centers, because grids are about sharing capacity and pooling resources.
Why did grid technology get its start in high energy physics? Because HEP runs on a scale that demands it, and because we have been dealing with sociological issues for a long time. If anyone can pave the way for grids, physics can. We have the arrogance and experience required to lead the way.
But grids cannot be just a throw-away technology, created for the moment and disposed of when the LHC has outrun its useful life.
The metric for success must be that the HEP community is no longer depended on as primary driver: that grid technology has taken on its own life and is self-perpetuating, and other scientific domains share in the burden of pushing it forward.
- Frank Würthwein, University of California San Diego