Section: New Results

Active and passive testing

Off-line test selection with test purposes for non-deterministic timed automata

Participants : Nathalie Bertrand, Thierry Jéron, Amélie Stainer.

In [17] , we propose novel off-line test generation techniques for non-deterministic timed automata with inputs and outputs (TAIOs) in the formal framework of the tioco conformance theory. In this context, a first problem is the determinization of TAIOs, which is necessary to foresee next enabled actions, but is in general impossible. The determinization problem is addressed in [18] thanks to an approximate determinization using a game approach (see 6.1.2.1 ). We adapt this procedure here to over- and under-approximation, in order to preserve tioco and guarantee the soundness of generated test cases. A second problem is test selection for which a precise description of timed behaviors to be tested is carried out by expressive test purposes modeled by a generalization of TAIOs. Finally, using a symbolic co-reachability analysis guided by the test purpose, test cases are generated in the form of TAIOs equipped with verdicts.

Test generation using pushdown automata

Participant : Puneet Bhateja.

IOLTS (input output labeled transition system) is a versatile model and is frequently used in model based testing to model the functional behavior of an IUT (implementation under test). However when a system is tested remotely, its observed behavior can be different from its actual functional behavior. In a previous paper, we defined a notion of remotely observed behavior of an IOLTS in terms of its actual behavior. Paper [14] contributes by proposing a methodology to simulate a PDA (pushdown automaton) from the given IOLTS such that the simulated PDA precisely expresses the remotely observed behavior of the IOLTS. The simulated PDA can be thought of as an automatic test generator for remote testing.

Test case selection in asynchronous testing

Participants : Puneet Bhateja, Thierry Jéron.

Conformance testing has a rich underlying formal theory called IOLTS-based conformance testing. Depending upon whether the implementation-under-test (IUT) interacts with its environment directly, or indirectly through a medium, IOLTS-based conformance testing can be classified as synchronous testing or asynchronous testing, respectively. So far the problem of test case selection has been addressed mostly in the context of synchronous testing. In this work we contribute by addressing this problem in the context of asynchronous testing. Though an asynchronously communicating process can be simulated by a synchronously communicating process, the fact that the simulating process is infinite state even if the simulated process is finite state made the problem challenging.

A tagging protocol for asynchronous testing

Participant : Puneet Bhateja.

Conformance testing has a rich underlying theory popularly called IOCO-test theory. In the realm of IOCO-test theory, this paper addresses the issue of testing a component of an asynchronously communicating distributed system. Testing a system which communicates asynchronously (i.e., through some medium) with its environment is more difficult than testing a system which communicates synchronously (i.e., directly without any medium). What impedes asynchronous testing is that the actual behavior of the implementation under test (IUT) appears distorted and infinite to the tester. This impediment consequently renders the problem of generating a complete test suite, from the given specification of the IUT, infeasible. To this end, paper [13] proposes a tagging protocol which when implemented by the asynchronously communicating distributed system will enable the generation of a complete test suite, from the specification of any of its component. Further, this paper describes how to generate the test suite from the given specification of the component.

Abstracting time and data for conformance testing of real-time systems

Participants : Thierry Jéron, Hervé Marchand.

Current approaches to model-based conformance testing of real-time systems are mostly based either on finite state machines/transition systems or on timed automata. However, most real-time systems manipulate data while being subjected to time constraints. The usual solution consists in enumerating data values (in finite domains) while treating time symbolically, thus leading to the classical state explosion problem. Paper [12] with W.L. Andrade and P. Machado (Fed. Univ. Campina Grande) proposes a new model of real-time systems as an extension of both symbolic transition systems and timed automata, in order to handle both data and time requirements symbolically. We then adapt the tioco conformance testing theory to deal with this model and describe a test case generation process based on a combination of symbolic execution and constraint solving for the data part and symbolic analysis for timed aspects.

Ensuring security properties

Runtime enforcement monitors: composition, synthesis, and enforcement abilities

Participant : Yliès Falcone.

Runtime enforcement is a powerful technique to ensure that a program will respect a given set of properties. In [9] we extend previous work on this topic in several directions. Firstly, we propose a generic notion of enforcement monitors based on a memory device and finite sets of control states and enforcement operations. Moreover, we specify their enforcement abilities w.r.t. the general Safety-Progress classification of properties. Furthermore, we propose a systematic technique to produce a monitor from the automaton recognizing a given safety, guarantee, obligation or response property. Finally, we show that this notion of enforcement monitors is more amenable to implementation and encompasses previous runtime enforcement mechanisms.

What can you verify and enforce at runtime?

Participant : Yliès Falcone.

The underlying property, its definition and representation play a major role when monitoring a system. Having a suitable and convenient framework to express properties is thus a concern for runtime analysis. It is desirable to delineate in this framework the sets of properties for which runtime analysis approaches can be applied to. [8] presents a unified view of runtime verification and enforcement of properties in the Safety-Progress classification. Firstly, we extend the Safety-Progress classification of properties in a runtime context. Secondly, we characterize the set of properties which can be verified (monitorable properties) and enforced (enforceable properties) at runtime. We propose in particular an alternative definition of ”property monitoring” to the one classically used in this context. Finally, for the delineated sets of properties, we define specialized verification and enforcement monitors.