Concept for the NASA risk and reliability data collection and analysis environment

The implementation of a NASA policy to collect R&M data to support ongoing and future projects has been limited, due in part to NASA's organization into largely independently operating programs and projects and the lack of a structured approach for the distribution of R&M data across those programs and projects. This paper describes the concept of a data collection and analysis environment that is being developed to support the collection, analysis, and distribution of R&M data. The purpose of the environment is to provide NASA projects with a standardized environment for storing information such as data sources, their interpretations, analyses and supplemental documentation. The environment is to provide an efficient way of collecting the information, and to allow for a controlled release to other projects at the program-, center-, or agency level. The data environment, which will largely be based on open-source software components, will allow raw data to be stored in a wide variety of electronic file formats, as well as in the form of links to other data repositories such as PRACA databases. The environment allows the relevant statistical information in those free-format data sources to be documented using predefined data types. Version control and workflow management capabilities will be provided, allowing proper configuration management processes to be applied to the data. The environment will embed a Bayesian data analysis capability, supported by a handbook describing solutions to commonly encountered risk and reliability parameter estimation problems, which was developed in a separately funded task. However, externally developed analyses can be stored inside the data environment as well, allowing configuration management and workflow management features to be taken advantage of. Results of such analyses may be encoded to make them available to reporting functions. The data environment is structured to allow projects to initially develop project-level data repositories, without exposing any information to other projects or programs. Once a mature state of the repository is reached, that state can be released to a central repository. This allows the information to be accessed by other projects.