Provenance in High-Energy Physics Workflows

The adoption of large-scale distributed computing for high-energy physics presents new opportunities and challenges for physicists analyzing the data from the Large Hadron Collider experiments. With petabytes of data to manage, accessed by thousands of systems and used by thousands of collaborators, effective provenance is critical to the understanding of how the physics results were produced. In this article, the authors discuss several uses of data provenance in high-energy physics workflows and the opportunities for improvements in data analysis workflows that result from decentralized provenance collection and fine-grained object annotations.