Designing Test Suites for Software Interactions Testing

Abstract

Testing is an expensive but essential part of any software project. Having the right methods to detect faults is a primary factor for success in the software industry. Component based systems are problematic because they are prone to unexpected interaction faults, yet these may be left undetected by traditional testing techniques. In all but the smallest of systems, it is not possible to test every component interaction. One can use a reduced test suite that guarantees to include a defined subset of interactions instead. A well studied combinatorial object, the covering array, can be used to achieve this goal. Constructing covering arrays for a specific software system is not always simple and the resulting object may not closely mirror the real test environment. Not only are new methods for building covering arrays needed, but new tools to support these are required as well. Our aim is to develop methods for building smaller test suites that provide stronger interaction coverage, while retaining the flexibility required in a practical test suite. We combine ideas from combinatorial design theory, computational search, statistical design of experiments and software engineering. We begin with a description of a framework for greedy algorithms that has formed the basis for several published methods and a widely used commercial tool. We compare this with a meta-heuristic search algorithm, simulated annealing. The results suggest that simulated annealing is more effective at finding smaller test suites, and in some cases improves on combinatorial methods as well. We then develop a mathematical model for variable strength interaction testing. This allows us to balance the cost and the time of testing by targeting individual subsets of components. We present construction techniques using meta-heuristic search and provide the first known bounds for objects of this type. We end by presenting some new cut-and-paste techniques that merge recursive combinatorial constructions

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2004
Accession Number
ADA628448

Entities

People

  • Myra B. Cohen

Organizations

  • University of Auckland

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies
  • Materials and Manufacturing Processes
  • Space

DTIC Thesaurus Topics

  • Algorithms
  • Application Software
  • Combinatorial Analysis
  • Communication Systems
  • Computational Science
  • Computer Programming
  • Computer Programs
  • Computers
  • Electronic Mail
  • Engineering
  • Experimental Design
  • Mathematical Models
  • Mathematics
  • Operating Systems
  • Software Development
  • Software Testing
  • Test Methods

Fields of Study

  • Computer science
  • Engineering

Readers

  • Operations Research
  • Sensor Fusion and Tracking Systems.
  • Systems Analysis and Design