Ludovic Henrio

Asynchronous and deterministic objects

This paper aims at providing confluence and determinism properties in concurrent processes, more specifically within the paradigm of object-oriented systems. Such results should allow one to program parallel and distributed applications that behave in a deterministic manner, even if they are distributed over local or wide area networks. For that purpose, an object calculus is proposed. Its key characteristics are asynchronous communications with futures, and sequential execution within each process.While most of previous works exhibit confluence properties only on specific programs -- or patterns of programs, a general condition for confluence is presented here. It is further put in practice to show the deterministic behavior of a typical example.

GCM: a grid extension to Fractal for autonomous distributed components

This article presents an extension of the Fractal component model targeted at programming applications to be run on computing grids: the grid component model (GCM). First, to address the problem of deployment of components on the grid, deployment strategies have been defined. Then, as grid applications often result from the composition of a lot of parallel (sometimes identical) components, composition mechanisms to support collective communications on a set of components are introduced. Finally, because of the constantly evolving environment and requirements for grid applications, the GCM defines a set of features intended to support component autonomicity. All these aspects are developed in this paper with the challenging objective to ease the programming of grid applications, while allowing GCM components to also be the unit of deployment and management.

Behavioural models for distributed Fractal components

This paper presents a formal behavioural specification framework for specifying and verifying the correct behaviour of distributed Fractal components. The first contribution is a parameterised and hierarchical behavioural model called pNets that serves as a low-level semantic framework for expressing the behaviour of various classes of distributed languages and as a common internal format for our tools. Then, we use this model to define the generation of behavioural models for applications ranging from sequential Fractal components, to distributed objects, and finally to distributed components. Our models are able to characterise both functional and non-functional behaviours and the interaction between the two concerns. Finally, this work has resulted in the development of tools allowing the non-expert programmer to specify the behaviour of his components and (semi)automatically verify properties of his application.

Behavioural models for hierarchical components

In this work, we focus on hierarchical component systems. We describe both the functional behaviour and the non-functional features (life-cycle management) of components in terms of synchronised transition systems; functional behaviours are supposed to be specified by the component developer, while management features can be built automatically for the architecture definition of a given component system. We define a notion of correct component composition; then we show how we can prove, using (compositional) model-checking techniques, temporal properties of a component system. Transformations of a system, for example replacement of a sub-component, are expressed as transformations of its behavioural semantics, allowing to prove preservation of some properties, or the validity of new properties after transformation.

A hybrid message Logging-CIC protocol for constrained checkpointability

Communication Induced Checkpointing protocols usually make the assumption that any process can be checkpointed at any time. We propose an alternative approach which releases the constraint of always checkpointable processes, without delaying any message reception nor altering message ordering enforced by the communication layer or by the application. This protocol has been implemented within ProActive, an open source Java middleware for asynchronous and distributed objects implementing the ASP (Asynchronous Sequential Processes) model.

Asynchronous sequential processes

Deterministic behavior for parallel and distributed computation is rather difficult to ensure. To reach that goal, many formal calculi, languages, and techniques with well-defined semantics have been proposed in the past. But none of them focused on an imperative object calculus with asynchronous communications and futures. In this article, an object calculus, Asynchronous Sequential Processes (ASP), is defined, with its semantics. We prove also confluence properties for the ASP calculus. ASPs main characteristics are asynchronous communications with futures, and sequential execution within each process. This paper provides a very general and dynamic property ensuring confluence. Further, more specific and static properties are derived. Additionally, we present a formalization of distributed components based on ASP, and show how such components are used to statically ensure determinacy. This paper can also be seen as a formalization of the concept of futures in a distributed object setting.

Verification of Distributed Hierarchical Components

Abstract Components allow to design applications in a modular way by enforcing a strong separation of concerns. In distributed systems this separation of concerns have to be composed with distribution of controls due to asynchrony. This article relies on Fractive, an implementation of the Fractal component model allowing to unify the notion of components with the notion of activity. This article shows how to build automatically the behaviour of a distributed component system. Starting from the functional specification of primitive components, we generate a specification of a system of components, their asynchronous communications, and their control. We then show how to use such a specification to verify properties specific to components, reconfigurations, or asynchrony.

Type Safe Algorithmic Skeletons

This paper addresses the issue of type safe algorithmic skeletons. From a theoretical perspective we contribute by: formally specifying a type system for algorithmic skeletons, and proving that the type system guarantees type safety. From an implementation point of view, we show how it is possible to enforce the type system on an Java based algorithmic skeleton library. The enforcement takes place at the composition of the skeleton program, by typing each skeleton with respect to its construction parameters: sequential functions, and other skeletons. As a result, hierarchical skeleton nesting can be performed safely, since type errors can be detected by the skeleton type system.

Functional Active Objects: Typing and Formalisation

This paper provides a sound foundation for autonomous objects communicating by remote method invo- cations and futures. As a distributed extension of @z-calculus, we define ASPfun, a calculus of functional objects, behaving autonomously and communicating by a request-reply mechanism: requests are method calls handled asynchronously and futures represent awaited results for requests. This results in a well structured distributed object language enabling a concise representation of asynchronous method invoca- tions. This paper first presents the ASPfun calculus and its semantics. Secondly we provide a type system for ASPfun, which guarantees the progress property. Most importantly, ASPfun and its properties have been formalised and proved using the Isabelle theorem prover, and we consider it as a good step toward formalisation of distributed languages.

A Fault Tolerant and Multi-Paradigm Grid Architecture for Time Constrained Problems. Application to Option Pricing in Finance.

This paper introduces a Grid software architecture offering fault tolerance, dynamic and aggressive load balancing and two complementary parallel programming paradigms. Experiments with financial applications on a real multi-site Grid assess this solution. This architecture has been designed to run industrial and financial applications, that are frequently time constrained and CPU consuming, feature both tightly and loosely coupled parallelism requiring generic programming paradigm, and adopt client-server business architecture.