[ Overview | Meetings | Visits | Events | Publications | Tools | Internal information ]
The objective of PROFUNDIS is to advance the state of the art of formal modelling and verification techniques to the point where key issues in mobile distributed systems, such as security protocols, authentication, access rights and resource management can be treated rigorously and with considerable automatic support. In particular we shall verify properties typical in so called open systems, where the behaviour of some parts (like intruders or adversaries) is unknowable, in extensible systems, where parts may be added or removed as the system executes, and in mobile systems where physical and logical connectivity between parts may change. We shall implement automatic and partly automatic analysis methods for ascertaining correct behaviour of such systems. For this purpose we shall integrate and focus several strands of ongoing theoretical work.
The work builds on recent advances in key theories for process behaviours, logics and types. We shall develop automata theoretic models suitable for our applications, with a particular interest in how they can be represented efficiently and used by automatic tools, and we shall determine how they are best used in connection with advanced forms of modal logics. The logics themselves will be developed, both in terms of their expressiveness for properties related to space and structure, and in terms of their accessibility and ease of use through suitable high-level representations.
We shall identify and develop analysis techniques related to these models and logics. This involves traditional behavioural equivalences and preorder checking, systematic simulation, and verification in interactive proof assistants. Here type systems will play an important role. Recent results show that types may themselves be used as crude but tractable correctness properties and therefore type inference is highly relevant, moreover, we shall explore how advanced type information can assist the other analysis techniques.
The ideas will to a large extent be implemented in a common tool set. Key issues here will be development and adaption of algorithms for analysis, and determining the best way of using them for practical examples. We shall in particular consider examples on security properties in systems for electronic commerce.
The results from the project will be of two kinds. First there will be a set of tools that automatically or with some manual assistance can treat the verification problems in the objectives above. The tools will be accompanied with a pragmatics, i.e., a guide for how they are effective for particular problems. Second the theories underpinning the tools will be demonstrated in scientific publications.