Overall Objectives
Research Program
Application Domains
New Software and Platforms
New Results
Partnerships and Cooperations
XML PDF e-pub
PDF e-Pub

Section: New Results

Foundations of Concurrency

Distributed systems have changed substantially in the recent past with the advent of phenomena like social networks and cloud computing. In the previous incarnation of distributed computing the emphasis was on consistency, fault tolerance, resource management and related topics; these were all characterized by interaction between processes. Research proceeded along two lines: the algorithmic side which dominated the Principles Of Distributed Computing conferences and the more process algebraic approach epitomized by CONCUR where the emphasis was on developing compositional reasoning principles. What marks the new era of distributed systems is an emphasis on managing access to information to a much greater degree than before.

A Concurrent Pattern Calculus

In [16] we detailed how Concurrent pattern calculus (CPC) drives interaction between processes by comparing data structures, just as sequential pattern calculus drives computation. By generalising from pattern matching to pattern unification, interaction becomes symmetrical, with information flowing in both directions. CPC provides a natural language to express trade where information exchange is pivotal to interaction. The unification allows some patterns to be more discriminating than others; hence, the behavioural theory must take this aspect into account, so that bisimulation becomes subject to compatibility of patterns. Many popular process calculi can be encoded in CPC; this allows for a gain in expressiveness, formalised through encodings.

An Intensional Concurrent Faithful Encoding of Turing Machines

The benchmark for computation is typically given as Turing computability; the ability for a computation to be performed by a Turing Machine. Many languages exploit (indirect) encodings of Turing Machines to demonstrate their ability to support arbitrary computation. However, these encodings are usually by simulating the entire Turing Machine within the language, or by encoding a language that does an encoding or simulation itself. This second category is typical for process calculi that show an encoding of lambda-calculus (often with restrictions) that in turn simulates a Turing Machine. Such approaches lead to indirect encodings of Turing Machines that are complex, unclear, and only weakly equivalent after computation. In [25] we developed an approach to encoding Turing Machines into intensional process calculi that is faithful, reduction preserving, and structurally equivalent. The encoding is demonstrated in a simple asymmetric concurrent pattern calculus before generalised to simplify infinite terms, and to show encodings into Concurrent Pattern Calculus and Psi Calculi.

Expressiveness via Intensionality and Concurrency

Computation can be considered by taking into account two dimensions: extensional versus intensional, and sequential versus concurrent. Traditionally sequential extensional computation can be captured by the lambda-calculus. However, recent work shows that there are more expressive intensional calculi such as SF-calculus. Traditionally process calculi capture computation by encoding the lambda-calculus, such as in the pi-calculus. Following this increased expressiveness via intensionality, other recent work has shown that concurrent pattern calculus is more expressive than pi-calculus. In [26] we formalised the relative expressiveness of all four of these calculi by placing them on a square whose edges are irreversible encodings. This square is representative of a more general result: that expressiveness increases with both intensionality and concurrency.

On the Expressiveness of Intensional Communication

The expressiveness of communication primitives has been explored in a common framework based on the pi-calculus by considering four features: synchronism (asynchronous vs synchronous), arity (monadic vs polyadic data), communication medium (shared dataspaces vs channel-based), and pattern-matching (binding to a name vs testing name equality). In [27] pattern-matching is generalised to account for terms with internal structure such as in recent calculi like Spi calculi, Concurrent Pattern Calculus and Psi calculi. This exploreD intensionality upon terms, in particular communication primitives that can match upon both names and structures. By means of possibility/impossibility of encodings, we showed that intensionality alone can encode synchronism, arity, communication-medium, and pattern-matching, yet no combination of these without intensionality can encode any intensional language.

Weak CCP Bisimilarity with Strong Procedures

Concurrent constraint programming (CCP) is a well-established model for concurrency that singles out the fundamental aspects of asynchronous systems whose agents (or processes) evolve by posting and querying (partial) information in a global medium. Bisimilarity is a standard behavioral equivalence in concurrency theory. However, only recently a well-behaved notion of bisimilarity for CCP, and a CCP partition refinement algorithm for deciding the strong version of this equivalence have been proposed. Weak bisimilarity is a central behavioral equivalence in process calculi and it is obtained from the strong case by taking into account only the actions that are observable in the system. Typically, the standard partition refinement can also be used for deciding weak bisimilarity simply by using Milner's reduction from weak to strong bisimilarity; a technique referred to as saturation. In [17] we demonstrated that, because of its involved labeled transitions, the above-mentioned saturation technique does not work for CCP. We gave an alternative reduction from weak CCP bisimilarity to the strong one that allows us to use the CCP partition refinement algorithm for deciding this equivalence.

Efficient Algorithms for Program Equivalence for Confluent Concurrent Constraint Programming

While the foundations and principles of CCP e.g., semantics, proof systems, axiomatizations, have been thoroughly studied for over the last two decades. In contrast, the development of algorithms and automatic verification procedures for CCP have hitherto been far too little considered. To the best of our knowledge there is only one existing verification algorithm for the standard notion of CCP program (observational) equivalence. In [18] we first showed that this verification algorithm has an exponential-time complexity even for programs from a representative sub-language of CCP; the summation-free fragment (CCP+). We then significantly improved on the complexity of this algorithm by providing two alternative polynomial-time decision procedures for CCP+ program equivalence. Each of these two procedures has an advantage over the other. One has a better time complexity. The other can be easily adapted for the full language of CCP to produce significant state space reductions. The relevance of both procedures derives from the importance of CCP+. This fragment, which has been the subject of many theoretical studies, has strong ties to first-order logic and an elegant denotational semantics, and it can be used to model real-world situations. Its most distinctive feature is that of confluence, a property we exploited to obtain our polynomial procedures.

A Behavioral Congruence for Concurrent Constraint Programming with Nondeterministic Choice

Weak bisimilarity is one of the most representative notions of behavioral equivalence for models of concurrency. As we mentioned earlier, a notion of weak bisimilarity, called weak saturated barbed bisimilarity (wsbb), was recently proposed for CCP. This equivalence improves on previous bisimilarity notions for CCP that were too discriminating and it is a congruence for the choice-free fragment of CCP. In [29] , however, we showed that wsbb is not a congruence for CCP with nondeterministic choice. We then introduced a new notion of bisimilarity, called weak full bisimilarity (wfb), and showed that it is a congruence for the full language of CCP. We also showed the adequacy of wfb by establishing that it coincides with the congruence induced by closing wsbb under all contexts. The advantage of the new definition is that, unlike the congruence induced by wsbb, it does not require quantifying over infinitely many contexts.

Abstract Interpretation of Temporal Concurrent Constraint Programs

Timed Concurrent Constraint Programming (tcc) is a declarative model for concurrency offering a logic for specifying reactive systems, i.e. systems that continuously interact with the environment. The universal tcc formalism (utcc) is an extension of tcc with the ability to express mobility. Here mobility is understood as communication of private names as typically done for mobile systems and security protocols. In [15] we considered the denotational semantics for tcc, and we extended it to a "collecting" semantics for utcc based on closure operators over sequences of constraints. Relying on this semantics, we formalized a general framework for data flow analyses of tcc and utcc programs by abstract interpretation techniques. The concrete and abstract semantics we proposed are compositional, thus allowing us to reduce the complexity of data flow analyses. We showed that our method is sound and parametric with respect to the abstract domain. Thus, different analyses can be performed by instantiating the framework. We illustrated how it is possible to reuse abstract domains previously defined for logic programming to perform, for instance, a groundness analysis for tcc programs. We showed the applicability of this analysis in the context of reactive systems. Furthermore, we made use of the abstract semantics to exhibit a secrecy flaw in a security protocol. We also showed how it is possible to make an analysis which may show that tcc programs are suspension free. This can be useful for several purposes, such as for optimizing compilation or for debugging.

Bisimulation for Markov Decision Processes through Families of Functional Expressions

In [24] , we transfered a notion of quantitative bisimilarity for labelled Markov processes to Markov decision processes with continuous state spaces. This notion takes the form of a pseudometric on the system states, cast in terms of the equivalence of a family of functional expressions evaluated on those states and interpreted as a real-valued modal logic. Our proof amounted to a slight modification of previous techniques used to prove equivalence with a fixed-point pseudometric on the state-space of a labelled Markov process and making heavy use of the Kantorovich probability metric. Indeed, we again demonstrated equivalence with a fixed-point pseudometric defined on Markov decision processes; what is novel is that we recasted this proof in terms of integral probability metrics defined through the family of functional expressions, shifting emphasis back to properties of such families. The hope is that a judicious choice of family might lead to something more computationally tractable than bisimilarity whilst maintaining its pleasing theoretical guarantees. Moreover, we used a trick from descriptive set theory to extend our results to MDPs with bounded measurable reward functions, dropping a previous continuity constraint on rewards and Markov kernels.