Abstract Details

<< Back to Schedule

9/28/2017  |   9:20 AM - 10:05 AM   |  

Data-Driven Management on a Large Software Intensive System: Insights and Lessons Learned

After decades in operation, NASA’s Space Network needed modernization. To do this NASA undertook the Space Ground Segment System Project. This is a large-scale, multi-year, software-intensive system development integrating a large number of COTS components and involving multiple sub-contractors. NASA is obviously highly vested in the progress and quality of the project, and has relied on the Fraunhofer USA Center for Experimental Software Engineering (FhUSA CESE) to provide independent 3rd party assessment of the projects performance and status. To provide NASA with the information and services it needs, Fraunhofer staff provide the program with broad engineering support, including providing software and system engineering development and technical expertise, assessing software artifacts, such as process documents, requirements and design documents, test plans and more, and supporting program management. In addition to serving as a research arm for the project, one way Fraunhofer supports NASA’s program management is by helping to track and forecast progress toward major milestones via Fraunhofer’s software process data analytics capabilities. Fraunhofer’s data-driven approach ultimately provides NASA with insights into the development progress and product quality, establishes baselines and assesses data against those baselines and other benchmarks, and enables management to quantitatively (rather than intuitively) manage the project. By providing NASA with greater visibility into the true state of the project, Fraunhofer objectively exposes project risks early. Our analysis and reporting are very well-received by NASA and are considered a useful input to their decision making. For instance, NASA has been able to nimbly make changes to the development process (e.g., re-planning, hiring of personnel, etc.), resulting in millions of dollars of cost savings. In the presentation, we will discuss our general approach toward measurement and data analysis, specifically as applied to the program. We will also discuss in more detail two analyses that we have performed for the program: Estimate to Completion and Defect Analysis. We will also share several best practices for establishing measurement programs to support project management.

This presentation has not yet been uploaded.

No handouts have been uploaded.

Madeline Diep (Primary Presenter), Fraunhofer USA Center for Experimental Software Engineering, mdiep@fc-md.umd.edu;
Dr. Madeline Diep is a Research Scientist at the Fraunhofer USA Center for Experimental Software Engineering in Maryland since 2009. At Fraunhofer, Dr. Diep is involved in a major software- intensive systems development effort for NASA, specifically in the analysis of system/software development metrics for monitoring and managing the development effort and risk. She has participated in a research on software and system assurance practices funded under NASA’s Office of and Safety Mission Assurance, developing best practices for formal inspections, threat modeling, and model-based software development. She has hands-on experience in building prediction models, and has provided measurement expertise to organizations implementing CMMI high maturity practice. She is working with a large organization to measure their data protection effectiveness. Her past projects include research and empirical studies in software instrumentation and monitoring for improving software testing and analysis. Her research interests are in the area of software quality and information assurance, testing, and software and process analytics. She taught Software Requirements course as the University of Maryland - College Park in 2009 through 2012. She holds a Ph. D in Computer Science from University of Nebraska - Lincoln.

2017 Sponsors: IEEE and IEEE Computer Society