Chander Sehgal, OSG project manager, wrote this column.
The Open Science Grid is a sharing ecosystem in computing for researchers in nearly all fields of science. OSG members connect their resources both to perform computing for their own experiments and to make available what they do not use at that time to other OSG members.
The collective nature of OSG makes science advances possible on a global scale. Many experiments have collaborators who are geographically scattered, and OSG allows them to connect their distributed computing resources to the “grid,” effectively using a common set of tools for that experiment.
Also, communities are often able to afford resources only for their average workloads, and that can create bottlenecks during peak periods such as during the run-up to summer conferences. With the sharing power of OSG, experiments can “borrow” computing opportunistically from other OSG partners to accomplish their work in time frames that would otherwise be impossible.
Over the last decade, Fermilab has been a vital partner in using and advancing OSG. CDF and DZero have used computing systems at Fermilab, such as FermiGrid, based on grid interfaces. CMS took it to the next level by building and operating the U.S. Tier 1 center, which serves as the U.S. hub for data and contributes a big share of the worldwide computing for CMS. And more recently, NOvA has started using OSG sites across the United States to run physics simulations.
Over the years, many Scientific Computing Division staff members have been key contributors to the advancement of OSG. Currently, Lothar Bauerdick serves as the executive director for the project, which executes the work program using staff from 11 U.S. institutions, and Ruth Pordes serves as the chair of the OSG Council, which represents the stakeholder community. In addition, Fermilab provides leadership in project management, user support and security for the OSG.
The 2014 OSG all-hands meeting at SLAC has just concluded. We reviewed the recent work of the project and learned about the use cases and experiences of many stakeholders scattered around the globe. This is an exciting period for OSG as we look to evolve and re-shape our platforms and services to continue to be an effective framework for distributed high-throughput computing for the next decade. The roots of OSG are in high-energy physics, but many non-physics researches now use it to study topics such as brain concussions and the human immune response to tuberculosis.
We are at an equally exciting time at Fermilab as we work with the Intensity Frontier experiments to enable their effective use of OSG and the distributed computing resources provided by Fermilab and other collaborators across the United States. The NOvA, MicroBooNE and LBNE collaborations are commissioning production activities on OSG now. We have a lot of hard work ahead as we build a distributed computing environment for the Intensity Frontier experiments, and I am confident in the future benefits for our researchers.