Building Containerized Environments for Reproducibility and Traceability of Scientific Workflows

TL;DR

This paper presents a containerized system that enhances reproducibility and traceability of scientific workflows by automatically capturing provenance metadata and building record trails with minimal overhead.

Contribution

It introduces a novel container-based environment that annotates workflows and captures provenance metadata, improving trust in simulation results.

Findings

Effective provenance metadata collection with low overhead

Built-in record trails improve transparency and reproducibility

Applicable to various scientific workflows

Abstract

Scientists rely on simulations to study natural phenomena. Trusting the simulation results is vital to develop sciences in any field. One approach to build trust is to ensure the reproducibility and traceability of the simulations through the annotation of executions at the system-level; by the generation of record trails of data moving through the simulation workflow. In this work, we present a system-level solution that leverages the intrinsic characteristics of containers (i.e., portability, isolation, encapsulation, and unique identifiers). Our solution consists of a containerized environment capable to annotate workflows, capture provenance metadata, and build record trails. We assess our environment on four different workflows and measure containerization costs in terms of time and space. Our solution, built with a tolerable time and space overhead, enables transparent and automatic provenance metadata collection and access, an easy-to-read record trail, and tight connections between data and metadata.