Skip to main content

Automated Verification of Infinite State Systems

Objective

Automatic verification of infinite state systems is a long-standing problem in Computer Science. The proposed project aims at designing a push-button technology, based on automated deduction, for error detection for infinite state systems including e-commerce and related security protocols. Similar to the paradigm shift achieved by adopting Ordered Binary Decision Diagrams in Symbolic Model Checking, we envisage that a major technological breakthrough in infinite state systems verification is possible by pushing emerging techniques in automated deduction. These techniques are on-the-fly symbolic exploration, constrained theorem-proving, and incremental state exploration via prepositional satisfiability checking. Each participant group is a leading expert in one of these domains: Deutsche Telekom, France Telecom and GEMPLUS have expressed their interest in partaking in the planned subsequent full RTD project if the one year assessment project is positive. Automatic verification of infinite state systems is a long-standing problem in Computer Science. The proposed project aims at designing a push-button technology, based on automated deduction, for error detection for infinite state systems including e-commerce and related security protocols. Similar to the paradigm shift achieved by adopting Ordered Binary Decision Diagrams in Symbolic Model Checking, we envisage that a major technological breakthrough in infinite state systems verification is possible by pushing emerging techniques in automated deduction. These techniques are on-the-fly symbolic exploration, constrained theorem-proving, and incremental state exploration via prepositional satisfiability checking. Each participant group is a leading expert in one of these domains: Deutsche Telekom, France Telecom and GEMPLUS have expressed their interest in partaking in the planned subsequent full RTD project if the one year assessment project is positive.

OBJECTIVES
This 1 year assessment project aims at laying the foundations of a new generation of verification tools for automatic error detection for e-commerce and related security protocols. To assess the potential of this technology, we will develop a prototype verification tool incorporating inference engines based on three promising automated deduction techniques: on-the-fly model-checking based on lazy data-types, theorem-proving with constraints, and model-checking based on prepositional satisfiability checking.
The assessment consists of two phases: a development phase aimed at the design and implementation of a prototype verification tool, and an analysis phase, in which the tool (and the techniques) will be tested and evaluated against a corpus of 50 security protocol verification problems. This will pave the way to turning the prototype into a mature technology, whose application in the industrial setting will be ascertained in a follow-up, full RTD project with industry involvement.

DESCRIPTION OF WORK
The project consists of two main tasks. To define a high-level language for specifying protocols, and design and implement a translator from protocol descriptions to a standard declarative format. To develop and test a technology for infinite state space exploration adapted to protocol verification and based on three techniques operating on the translator's output. The first technique, on-the-fly model checking, uses lazy data-types and specialized algorithms that can automatically handle infinite state spaces. The second technique, theorem-proving with constraints, provides an efficient way of representing an infinite state space using a constraint store. Additionally, it offers advantages in checking timing and freshness properties, which are crucial for security protocols. In both these techniques, flaws are detected by efficient pattern matching on traces. Often infinite state spaces can be iteratively approximated by large finite states spaces. The third technique will employ model-checking techniques based on propositional satisfiability checking to reason about these approximations. Although each technique can work independently, they will be integrated into a single prototype verification tool where they will interact and benefit from each other's strengths. This will require foundational research in the scope and limitations of our symbolic reasoning techniques (completeness of simplifications, complexity, expressiveness) as well as advances in integrating cooperating semi-decision procedures. In parallel to the above activities, collaboration with industrial partners will be initiated in order to identify a set of representative case studies coming from the industrial practice on which to apply the results of the project.
The project has broadly fulfilled its objectives:
1. The land surface schemes of two regional models, two climate models and one model of intermediate complexity have been tested against data sets from EFEDA and HAPEX. Improvements have been made where necessary.
2. Data sets of current and past land cover have been accessed. Future scenarios have been designed - including the use of a socio-economic model in the Sahelian region to define past, current and future land use patterns.
3. A variety of experiments have been undertaken to investigate the impact on the climate of sea surface temperature fields, intensity of land degradation, the spatial extent of land degradation, interactive vegetation. Significant improvements to the operation of the models have been achieved - leading to an increase in realism of the simulations. Significant improvements in our understanding of the role of the land surface in climate have been realised.

Funding Scheme

ACM - Preparatory, accompanying and support measures

Coordinator

ALBERT-LUDWIGS-UNIVERSITAET FREIBURG
Address
Fahnenbergplatz
79085 Freiburg
Germany

Participants (2)

INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET EN AUTOMATIQUE
France
Address
Domaine De Voluceau
78153 Le Chesnay
UNIVERSITA DEGLI STUDI DI GENOVA
Italy
Address
Via Balbi 5
16126 Genova