Testing IoT Systems

Published: March 26, 2018

Author(s)

Jeffrey M. Voas, David R. Kuhn, Phil Laplante

Abstract

The ability to test systems that are based on the underlying products and services commonly referred to as the Internet of 'things' (IoT) is discussed. The role of a static metric that can be applied to design, architectures, hardware, 'things', and software is discussed. The metric, the Domain Range Ratio (DRR) [2], is simply the number of test cases divided by the number of system outputs. Any system built from IoT products and services, termed a Network of Things (NoT) [1], will likely have low DRRs. Low DRRs hint that it will be harder for test cases to detect bugs since internal state corruptions will not propagate and therefore remain undetected. Low DRR scores also suggest where to place internal test assertions. Combinatorial testing is a technique that offers test case minimization yet without sacrificing fault detection. In this paper, the level of combinatorial testing as a function of DRR scores is studied, leading to improved fault detection for a specific NoT during test.
Conference Dates: March 26-29, 2018
Conference Location: Bamberg, -1
Conference Title: 12th IEEE International Symposium on Service-Oriented System Engineering
Pub Type: Conferences

Keywords

Internet of Things (IoT), Network of Things (NoT), software testing, combinatorial testing, Domain Range Ratio (DRR), assertions, test case reduction
Created March 26, 2018, Updated August 08, 2018