FSM-Based Test Derivation Strategies for Systems with Time-Outs

Experimental Comparison of the Quality of TFSM-Based Test Suites for the UML Diagrams

“…The second method (Method 2) for deriving a test suite against TFSMs with the guaranteed fault coverage is based on the correlation between TFSM and FSM (Procedure 1) [4]. To transform a timed FSM into a classical FSM we add a special input symbol 1 that corresponds to the notion of waiting one time unit, and a special output -N that corresponds to the case when there is no reply from the machine.…”

“…Consider now the third test derivation method proposed in the paper [4]. The method has two testing assumptions: the upper bound on the number of states of a TFSM under test (implementation under test, IUT) and the largest finite timeout at a state of the IUT are known.…”

“…In this paper, we derive tests with the guaranteed fault coverage based on a timed Finite State Machine (FSM) augmented with timeouts [1], since FSMs are known to be an efficient model for deriving tests with the guaranteed fault coverage. The paper presents a case study for assessing the quality of test suites derived by three methods [2,3,4], which is estimated for the example of a phone line; the UML description of this project is taken from [5]. Using the tool Visual Paradigm for UML 8.0 [6] a JAVA code is generated for this application that serves a sample when assessing the test suite quality.…”

Detecting Faults in TFTP Implementations using Finite State Machines with Timeouts

“…In order to describe the TFSM behavior we use the notion of a timed input that indicates that an input i is applied at time t. Correspondingly a timed input is a pair (i, t)  I  Z + where Z + is the set of nonnegative integers. In order to extend the transition relation to timed inputs we have to know at which state s is an FSM when applying an input i at time t and the state s is determined based on the timeout function [12]. Given a TFSM S, state s of the TFSM is an input-reachable state (ir-state) if there exists a timed trace (, ) such that (s 0 , , , s)   S [12].…”

“…Correspondingly, the behavior of an FMS significantly depends on a time instance when an input is applied, i.e., the behavior of the FSM is specified for timed input sequences. In [12], it is shown how this behavior can be described by an ordinary FSM with an additional input symbol (a time unit) and thus, despite of the fact that a test suite derived for such an abstract FSM using black-box testing methods returns highly redundant tests, FSM-based test derivation methods can be directly used when deriving tests from an FSM with timeouts. In this paper, we derive a test suite as a transition tour of an appropriate FSM, since W-based testing methods [3] ask for the specification FSM to be complete and deterministic and this usually does not hold for FSMs which describe protocol behavior.…”