IDENTIFICATION AND SIGNIFICANCE OF INNOVATION
In recent years, industry and government have begun adopting Model Based Engineering (MBE) practices in an unprecedented way. No longer exclusively the domain of isolated experts, MBE has been implemented across the full product lifecycle. This success is, however, giving rise to new challenges. The ability to share disparate models across teams, organizational boundaries, and among communities of practice is important for collaboration on complex projects. Model reuse is important because it minimizes the need to “re-invent the wheel” for each new project or initiative. Finally, the need to rapidly integrate disparate models together is critical for accurate evaluation of multidisciplinary systems and for evaluating different system/mission concepts and architectural variants. Phoenix Integration proposes to address these challenges by developing an analysis model sharing platform similar to GitHub, a successful software development collaboration system. This platform will be easy-to-use, web-based, and built on the Git version control system. Modifications will be made to tailor the Git repository system to version control analysis models and workflows. This model-sharing platform will have provision for documentation, tags and metadata, which includes a software dependencies list and supported OS. Shared analyses and workflows would be easily run in ModelCenterreg; and Jupyter notebooks. It will rely on cloud computing resources with on-demand provisioning and execute the analyses and workflows on them when requested. Additionally, published analyses and workflows would be verified automatically whenever supporting software versions change.
TECHNICAL OBJECTIVES AND PROPOSED DELIVERABLES
The overarching objective of this project is the prototype development and demonstration of an accessible easy-to-use web-based model sharing platform coupled with a consistent and repeatable analysis execution platform. This will be done by leveraging existing and current technology already proven in the marketplace. The Phase II objectives are to prototype and demonstrate the following:
- The use of the Git version control system to house ModelCenter® analyses and workflows
- Use software containerization to reliably reproduce the software environment for workflow execution
- Provide a methodology to make software containers built for workflow execution easily and readily available to an engineering community
- Leverage the automation of available computational resources to automatically execute workflows
- The creation of a web-based portal where workflow repositories can be accessed with provision for documentation, tags and metadata.
- Integrate with PLM and SPDM software to allow the consistent storage of workflow data products, together with any necessary data for reproducibility and traceability
- Deploy prototype implementations to NASA and industry partners as soon as available to solicit valuable feedback and comments
A successful project will help NASA further advance the MBE vision and will help enable more comprehensive, broader, and deeper modeling efforts across all of NASA’s programs. Specifically, the project will help NASA to share engineering models and workflows across teams, organizational boundaries, and among communities of practice. It will also enable model traceability, repeatability and reusability. These capabilities will directly benefit ongoing and future NASA projects and initiatives, such as the Mars 2020 and Europa Clipper missions.
We propose developing an easy-to-use analysis sharing platform, while enabling reliable model executions. Executing these analyses on cloud computing resources opens the possibility of easy accessibility, including automatic model verification. This can be used across all areas of engineering, including those in the aerospace & defense, automotive, scientific research and heavy industries.