|
|
|
| e-Pub |
Previous | 
Home
Section: New Results
Parallel and Distributed Verification
Distributed State Space Manipulation
Participants : Hubert Garavel, Hugues Evrard, Wendelin Serwe.
For distributed verification, CADP provides the PBG format, which implements the theoretical concept of Partitioned LTS [38] and provides a unified access to an LTS distributed over a set of remote machines.
In 2016, the code of the CAESAR_NETWORK_1 library, which is a building block for the distributed verification tools of CADP, has been carefully scrutinized and split into logically-independent modules. Nine problems have been detected and solved, among which a flaw in the distributed termination algorithm: the entire network could freeze if a worker process crashed too early, before the opening of TCP sockets. From now on, a better distributed termination algorithm is used, which supports the coexistence of several networks, ensures that all connections are closed before terminating, and produces more informative traces indicating which worker has triggered termination. Also, the improved CAESAR_NETWORK_1 library checks that all workers operate in directories that are pairwise distinct, mutually disjoint, and different from the working directory of the coordinator process.
Distributed Code Generation for LNT
Participants : Hugues Evrard, Frédéric Lang, Wendelin Serwe.
Rigorous development and prototyping of a distributed algorithm using LNT involves the automatic generation of a distributed implementation. For the latter, a protocol realizing process synchronization is required. As far as possible, this protocol must itself be distributed, so as to avoid the bottleneck that would inevitably arise if a unique process would have to manage all synchronizations in the system. Using a synchronization protocol that we verified formally in 2013, we developed a prototype distributed code generator, named DLC (Distributed LNT Compiler), which takes as input the model of a distributed system described as a parallel composition of LNT processes.
In 2016, we improved the user documentation of the DLC distribution, and added support for structured data types, enabling experiments of DLC on the LNT model of the CAESAR_SOLVE_2 library (see § 6.2.3). An overview paper about DLC has been accepted in an international journal [12].
Distributed Resolution of Boolean Equation Systems
Participant : Wendelin Serwe.
The BES_SOLVE tool of CADP enables to solve BESs (Boolean Equation Systems) using the various resolution algorithms provided by the CAESAR_SOLVE library (see 5.1), including a distributed on-the-fly resolution algorithm described in pseudo-code in [45].
In 2016, we modeled the pseudo-code of the distributed resolution algorithm in LNT (about lines). For a set of BES examples (encoded as LNT data types and functions), we experimented the generation of the LTS corresponding to the distributed resolution algorithm applied to each BES. We also experimented with the DLC tool [21] to generate a prototype distributed implementation of the resolution algorithm from its LNT specification. These experiments uncovered some errors in the original pseudo-code.
We also simplified the C implementation included in the BES_SOLVE tool to closer match the corrected LNT model, mainly by removing additional synchronization messages. We started to evaluate the simplified implementation using our non-regression test base (more than BESs), with promising results.
Stability of Communicating Systems
Participant : Gwen Salaün.
Analyzing systems communicating asynchronously via reliable FIFO buffers is an undecidable problem. A typical approach is to check whether the system is bounded, and if not, the corresponding state space can be made finite by limiting the presence of communication cycles in behavioral models or by imposing an upper bound for the size of communication buffers.
In 2016, our focus was on systems that are likely to be unbounded and therefore result in infinite systems. We did not want to restrict the system by imposing any arbitrary bound. We introduced a notion of stability and proved that once the system is stable for a specific buffer bound, it remains stable whatever larger bounds are chosen for buffers. This enables one to check certain properties on the system for that bound and to ensure that the system will preserve them for arbitrarily larger buffer bounds. We also proved that computing this bound is undecidable but we showed how we succeed in computing these bounds for many examples using heuristics and equivalence checking. These results have been published in an international conference [18].
In collaboration with Carlos Canal (University of Málaga, Spain), we have also shown how the stability approach can be used for composition and adaptation of component-based software. This led to a publication in an international conference [20].
Debugging of Concurrent Systems
Participants : Gianluca Barbon, Gwen Salaün.
Model checking is an established technique for automatically verifying that a model satisfies a given temporal property. When the model violates the property, the model checker returns a counterexample, which is a sequence of actions leading to a state where the property is not satisfied. Understanding this counterexample for debugging the specification is a complicated task for several reasons: (i) the counterexample can contain hundreds of actions, (ii) the debugging task is mostly achieved manually, and (iii) the counterexample does not give any clue on the state of the system (e.g., parallelism or data expressions) when the error occurs.
In 2016, we proposed a new approach that improves the usability of model checking by simplifying the comprehension of counterexamples. Our solution aims at keeping only actions in counterexamples that are relevant for debugging purposes. To do so, we first extract in the model all the counterexamples. Second, we define an analysis algorithm that identifies actions that makes the behaviour skip from incorrect to correct behaviours, making these actions relevant from a debugging perspective. Our approach is fully automated by a tool that we implemented and applied on real-world case studies from various application areas for evaluation purposes. A paper presenting these results has been accepted at an international conference.