About This Document#
DesignSafe provides access to powerful computational resources through TACC, but using those resources effectively requires understanding how computation, data, and interfaces fit together. This document is designed to provide that foundation.
Rather than focusing on a single application, this book introduces the computational workflow architecture of DesignSafe — explaining how analyses move from user-facing interfaces, through middleware services, onto compute systems, and back again as results. The emphasis is on how the platform works as a system, and how researchers can design workflows that scale from interactive exploration to large, automated studies.
How This Book Is Organized#
This book is organized into several complementary layers:
Computational workflow concepts Introduces common workflow patterns (interactive, batch, automated) and the architectural components that support them.
Compute environments Describes the execution environments available on DesignSafe — including JupyterHub, HPC systems, and virtual machines — and explains when and why each is appropriate.
SLURM and batch execution Provides a scheduler-level view of how jobs are submitted, scheduled, and executed on TACC systems, forming the basis for scalable computation.
Tapis and Tapis Apps Introduces the middleware layer that connects user interfaces to compute resources, including job submission, monitoring, automation, and application abstraction.
Python utilities and training notebooks Supplies reusable tools and hands-on notebooks that demonstrate how these concepts are applied in practice and how workflows can be automated programmatically.
File storage and data movement Explains how data is staged, accessed, moved, and persisted across systems, which is essential for building reliable workflows.
These sections are designed to be modular. While the book can be read sequentially, many chapters also serve as reference material that users may return to as their workflows evolve.
Relationship to Application-Specific Documents#
This document establishes a general framework for computational research workflows on DesignSafe. Application-focused documents (such as those for OpenSees, OpenFOAM, or ADCIRC) build on this foundation by applying the same concepts to specific tools.
Those documents are intended as worked examples, not prerequisites. Understanding the workflow architecture presented here will make it easier to adapt to new applications and computational methods over time.
Python Utilities Library Used in This Document#
Throughout the notebooks and examples in this book, a custom Python function library is used to support common workflow tasks such as job submission, monitoring, file management, and data inspection. This library is stored in DesignSafe Community Data and is actively maintained.
While the library can be copied directly, this is not recommended, as it is updated regularly.
Each notebook includes a small setup snippet near the top showing how to reference or copy a stable version of the library into your own workspace. This approach balances reproducibility with ongoing improvements.