J. Wu, L. Xu, and C. Liu, "Develop a Reliability Test in TTCN-3," ETSI TTCN-3 User Conference 2007: Stockholm (Sweden), May 29-June 1, 2007.
Added by: Deleted user (01 Aug 2008 14:40:05 Europe/Berlin) Last edited by: Deleted user (20 Aug 2008 12:19:56 Europe/Berlin)
|Resource type: Conference Paper
BibTeX citation key: Wua
View all bibliographic details
Creators: Liu, Wu, Xu
Publisher: Ericsson, ETSI (Stockholm (Sweden))
Collection: ETSI TTCN-3 User Conference 2007
Views index: %
Popularity index: 10.25%
TTCN-3 (Testing and Test Control Notation-version 3) is a standardized test specification and implementation language. It has been widely used in industrial testing. The development of test system in TTCN-3 is more concentrated on functional testing so far, relatively few on non-functional testing, such as reliability testing.
When using TTCN-3 to specify a functional test, testers usually define test cases in the enumerative way without expressing what the SUT is. This methodology, when applied to develop reliability test, would suffer from the following problems:
(1) Test cases for reliability testing are usually generated from usage model by random sampling according to a certain distribution. The number of test cases required is usually too large to be defined by manual enumeration in TTCN-3;
(2) It requires re-sampling the test cases from usage model in each reliability growth test because a fixed set of test cases cannot reflect the true reliability growth. This requirement makes defining test cases manually impossible;
(3) Reliability measurement is a necessary step in reliability testing, which requires the usage model as source data. Because no way to specify usage model in TTCN-3 test system, tester has to collect data manually from multiple places.
Our research objective is to propose a solution that could automatically generate the reliability test cases in TTCN-3. The keys to develop such solution lie on how to specify usage model, how to connect usage model with TTCN-3 test behaviour, and how to generate TTCN-3 test case from usage model. We develop a TTCN-3 test composing tool which takes the usage model (in xml) and SUT stimulus (in TTCN-3) as input. The composing tool generates test cases automatically in three steps:
(1) Samples the usage model to get conceptual test cases;
(2) Rewrites the conceptual test cases in TTCN-3 stimulus;
(3) Composes and outputs the TTCN-3 test system.
To evaluate the effectiveness of the solution, we conduct an experiment on Web application reliability testing. The usage model is specified as Markov Chain Usage Model (MCUM) in which node is modelled as SUT usage state, transition is modelled as stimuli mapped to the stimuli in TTCN-3. Based on the experiment, this report also discusses the possible extensions for TTCN-3 when applied to develop reliability test.
Added by: Deleted user