Digital Engineering-driven Software Quality Assurance-as-a-Service  
Author Kazuhira Okumoto

 
Co-Author(s)

 
Abstract Recent advances in technology are pushing the limits in computation, networking, storage, and are allowing an increasing number of business processes to be intelligently automated. These advances require, in a large part, the ability to continuously design, develop, test, integrate and deploy high quality software. More than ever, software development teams need ways to plan their release timelines, and to ensure that released software is of the expected quality. While software quality assurance models are well-studied, there remains a significant gap in translating these models to practice. In practice, subject matter experts are required to continuously tweak models and their parameters, and generally perform software reliability analysis using spreadsheets. This usually generates inconsistent and biased results. The results are not easily shared within the software development team, which creates a lack of communication. In this paper, we present STAR, a novel cloud-based tool for software quality assurance that is driven by digital engineering principles. The core engine of STAR implements a series of piece-wise exponential models for defect trend analysis. In addition, during early stages of projects where there is not defect data, STAR proposes an algorithm that uses software development effort plans from previous releases to predict defects for a current release. Furthermore, STAR includes prescriptive scenario planning, which allows users to run interactive whatif scenarios and corrective actions needed to meet delivery and quality targets. All these have been implemented in a software-asa- service model that allows users to create projects and releases, and automatically obtain all the necessary quality metrics to decide whether the release is ready for high-quality delivery. Evaluations of the effectiveness of STAR have been performed using data from large-scale software development projects, as well as through collaborations with large industry and academia partners.

 
Keywords Software reliability, digital engineering, software quality, on-time delivery, enterprise and program decisionmaking, program readiness, prescriptive data analytics, engineering practice, decision analytics, and visualization
   
    Article #:  RQD28-15
 

Proceedings of 28th ISSAT International Conference on Reliability & Quality in Design
August 3-5, 2023