Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
A survey of tool-supported assurance case assessment techniques
Maksimov M., Kokaly S., Chechik M. ACM Computing Surveys52 (5):1-34,2019.Type:Article
Date Reviewed: Mar 1 2021

Assurance case assessment techniques find application in the study of complex system reliability and safety, for example, nuclear reactors, energy systems, aircraft, and so on.

As the authors state, “significant research has gone into developing and surveying methods for evaluating/assessing assurance cases.” The present article is based on a survey of published works, from 1998 and later; ten papers were selected out of 37 for comparative analysis. The published literature identifies the following commercial/open-source assessment tools: ACedit, ASCE, AutoFOCUS 3, CertWare, D-Case Editor, NOR-STA, and SafeEd. The survey aims to assess assurance cases, assurance case tools, and the connections between them.

The assessment tools were then classified based on support for assurance case structural assessment: “Are structural constraints during argument creation available? Can the model be validated for correctness/completeness? Are syntactic queries ... allowed?” The probabilistic models used either Bayesian or Dempster-Shafer theory, or continuous/discrete-event simulation; the majority preferred Dempster-Shafer theory.

The survey targets five types of assessment: argument, evidence, tracing, reporting, and interaction; these were met by seven of the ten surveyed tools. The assurance process aims to arrive at the goal/claim to be justified, which represent the root node of a tree structure. The supporting elements of the tree construct assurance cases, and their parents represent an overall claim made about subsystems. A goal is decomposed into a number of progressively more specific subgoals, which lead to solution nodes representing evidence that justifies theclaim. Further, confidence values are assigned to nodes of the tree and link strengths are assigned weights.

Tools that base their argument assessment on Bayesian reasoning and leaf nodes are manually assigned prior probabilities based on available evidence, and the probabilities of claims are computed using Bayesian inferences. Some tools are built on logic programming, and the assurance case created is in the form of a proof, while some use plugins to create extensions.

It is found that a majority of the assessment techniques heavily rely on manual efforts, particularly when supporting evidence is in the form of natural language, and focus on the assessor’s ability “to convey his/her judgment in the form of quality and confidence values.” Furthermore, the study finds that “no regulatory bodies currently consider these assessment features in the software certification process.”

In summary, the authors present a thorough survey on tool support for assurance case assessment, and specifically a novel study of a structural checks approach methodology for such tools.

Reviewer:  K R Chowdhary Review #: CR147201 (2105-0113)
Bookmark and Share
  Featured Reviewer  
 
Introductory And Survey (A.1 )
 
 
Reference (A.2 )
 
Would you recommend this review?
yes
no
Other reviews under "Introductory And Survey": Date
Computer science: an overview (3rd ed.)
Brookshear J., Benjamin-Cummings Publ. Co., Inc., Redwood City, CA, 1991. Type: Book (9780805302646)
Feb 1 1992
Computers and applications with Basic
Slotnick D., Butterfield E., Colantonio E., Kopetzky D., Slotnick J., D. C. Heath and Company, Lexington, MA, 1990. Type: Book (9780669217537)
Dec 1 1991
Computers
Wear L., Pinkert J., Wear L., Lane W., McGraw-Hill, Inc., New York, NY, 1991. Type: Book (9780070686748)
Nov 1 1991
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy