TRACE: TRAnsparency CErtified
Building trust in computational research through certification
Project Overview
The TRACE (TRAnsparency CErtified) project is developing a new approach to computational transparency and reproducibility in research. Rather than requiring consumers of research to repeat computations to verify results, TRACE enables trust through certifying the successful original execution of a computational workflow that produced findings in situ.
The Problem
Research communities across the sciences increasingly require that authors share the data, code, and methods used to obtain computational results. However, without verification by repeating computations, there is no guarantee that provided artifacts are complete or can actually be used to produce reported results.
Particularly problematic are studies that employ:
- Sensitive or proprietary data for which access and reuse are restricted
- Streaming, transient, or ephemeral data that cannot be used to verify reproducibility due to their dynamic nature
- Very large-scale or specialized computational resources available only to authorized users
In these cases, verification by repeating computations may not be possible.
The Solution
TRACE is developing a model of certified transparency whereby the original execution of a computational workflow is certified by the system on which it was run. With certifications in hand, consumers of research can trust the transparency of results without necessarily repeating computations.
Key Components
The TRACE project includes:
- Technical specifications for describing and certifying research artifacts
- Reusable software tools for implementing certification systems
- Integration with continuous integration (CI) systems (GitHub Actions, Bitbucket Pipelines, Azure)
- Support for HPC environments (SLURM clusters and other high-performance computing systems)
Funding
This material is based upon work supported by the National Science Foundation under:
- Award Number: OAC-2209628 (University of Illinois)
- Award Number: OAC-2209629 (Cornell University)
- Award Number: OAC-2209630 (University of North Carolina)
Collaborating Institutions
- University of Illinois at Urbana-Champaign
- Cornell University
- University of North Carolina at Chapel Hill