Philippe Massonet

Agent Oriented Analysis Using Message/UML

This paper presents the MESSAGE/UML agent oriented software engineering methodology and illustrates it on an analysis case study. The methodology covers MAS analysis and design and is intended for use in mainstream software engineering departments. MESSAGE integrates into a coherent AOSE methodology some basic agent related concepts such as Organization, role, goal and task, that have so far been studied in isolation. The MESSAGE notation extends the UML with agent knowledge level concepts, and diagrams with notations for viewing them. The proposed diagrams extend UML class and activity diagrams.

Reservoir - When One Cloud Is Not Enough

As cloud computing becomes more predominant, the problem of scalability has become critical for cloud computing providers. The cloud paradigm is attractive because it offers a dramatic reduction in capital and operation expenses for consumers.

GRAIL/KAOS: an environment for goal-driven requirements engineering

The KAOS methodology provides a language and method for goal-driven requirements elaboration. GRAIL is an environment under development to support the KAOS methodology. The GRAIL kernel combines a graphical view, a textual view, an abstract syntax view, and an object base view of specifications. GRAIL has been used to elicit and specify the requirements of several real, industrial projects.

Goal-oriented requirements animation

Requirements engineers need to make sure that the requirements models and specifications they are building do accurately capture what stakeholders really want. Requirements animation has been recognized to be a promising approach to support this. The principle is to simulate an executable version of the requirements model and to visualize the simulation in some form appealing to stakeholders. Most animation tools available to date simulate operational models. Such models in general do not directly reflect the objectives, constraints and assumptions stated declaratively by stakeholders. It is also not possible to focus the animation on particular portions of a complex model relevant to some specific concern. The paper describes a tool aimed at overcoming such limitations by animating goal-oriented requirements models. The tool automatically generates parallel state machines from goal operationalizations, instantiates those machines to specific instances created by users at animation time, executes them from concurrent events input by multiple users, monitors property violations at animation time, and visualizes concurrent simulations in terms of animated scenes in the domain.

Early Verification and Validation of Mission Critical Systems

Our world is increasingly relying on complex software and systems. In a growing number of fields such as transportation, finance, telecommunications, medical devices, they now play a critical role and require high assurance. To achieve this, it is imperative to produce high quality requirements. The KAOS goal-oriented requirements engineering methodology provides a rich framework for requirements elicitation and management of such systems. This paper demonstrates the practical industrial application of that methodology. The non-critical parts are modelled semi-formally using a graphical language for goal-oriented requirements engineering. When and where needed (ie. for critical parts of a system) the model can be specified at formal level using a real-time temporal logic. That formal level seamlessly extends the semi-formal level which can also help hide the formality for the non-specialist. To ensure at an early stage that the right system is being built and that the requirements model is right, validation and verification tools are applied on that model. Early verification checks help to discover missing requirements, overlooked assumptions or incorrect goal refinements. State machines generation from operations provides an executable model useful for validation purposes or for deriving an initial design. Acceptance test cases and runtime behavior monitors can also be derived from the model. The process is supported by an integrated toolbox implementing the above tools by a roundtrip mapping of KAOS requirements level notations to the languages of formal technology tools such as model-checkers, SAT engines or constraint solvers. A graphical visualization framework also significantly helps validation using domain-based representations.

A Monitoring and Audit Logging Architecture for Data Location Compliance in Federated Cloud Infrastructures

Current cloud infrastructures have opaque service offerings where customers cannot monitor the underlying physical infrastructure. This situation raises concerns for meeting compliance obligations by critical business applications with data location constraints that are deployed in a Cloud. When federated cloud infrastructures span across different countries where data can migrate from one country to another, it should be possible for data owners to monitor the location of their data. This paper shows how an existing federated Cloud monitoring infrastructure can be used for data location monitoring without compromising Cloud isolation. In the proposed approach collaboration is required between the cloud infrastructure provider (IP) and the user of the cloud, the service provider (SP): the IP monitors the virtual machines (VM) on the SPs behalf and makes the infrastructure level monitoring information available to him. With the monitoring information the SP can create the audit logs required for compliance auditing. The proposed logging architecture is validated by an e-Government case study with legal data location constraints.

Early verification and validation of mission critical systems

Complex software and systems are pervasive in today’s world. In a growing number of fields they come to play a critical role. In order to provide a high assurance level, verification and validation (V&V) should be considered early in the development process. This paper shows how this can be achieved based on a goal-oriented requirements engineering framework which combines complementary semi-formal and formal notations. This allows the analyst to formalize only when and where needed and also preserves optimal communication with stakeholders and developers.For the industrial application of the methodology, a supporting toolbox was developed. It consist of a number of tightly integrated tools for performing V&V tasks at requirements level. This is achieved through the use of (1) a roundtrip mapping between the requirements language and the specific formal languages used in the underlying formal tools (such as SAT or constraint solvers) and (2) graphical views using domain-based representations. This paper will focus on two major and representative tools: the Refinement Checker (about verification) and the Animator (about validation).

Towards Modelling Obligations in Event-B

We propose a syntactic extension of Event-B incorporating a limited notion of obligation described by triggers. The trigger of an event is the dual of the guard: when a guard is not true, an event must not occur, whereas when a trigger is true, the event must occur. The obligation imposed by a trigger is interpreted as a constraint on when the other events are permitted. For example, the simplest trigger next, which states that the event must be the next one to be executed when the trigger becomes true, is modelled as an extra guard on each of the other events which prohibits their execution at this time. In this paper we describe the modelling of triggers in Event-B, and analyse refinement and abstract scheduling of triggered events.

Management and service-aware networking architectures (MANA) for future Internet — Position paper: System functions, capabilities and requirements

Future Internet (FI) research and development threads have recently been gaining momentum all over the world and as such the international race to create a new generation Internet is in full swing: GENI [16], Asia Future Internet [19], Future Internet Forum Korea [18], European Union Future Internet Assembly (FIA) [8]. This is a position paper identifying the research orientation with a time horizon of 10 years, together with the key challenges for the capabilities in the Management and Service-aware Networking Architectures (MANA) part of the Future Internet (FI) allowing for parallel and federated Internet(s).

Agents and the UML: A Unified Notation for Agents and Multi-agent Systems?

Over the last few years, agent-oriented software engineering has promoted the adoption of agents as a first-class paradigm for software engineering in research and industrial development. Agents have been used in research development for more than twenty years, while they still do not find complete acceptance in industrial settings. We believe that basically three characteristics of industrial development prevent the adoption of agents: (i) The scope of industrial projects is much larger than typical research efforts, (ii) The skills of developers are focused on established technologies, (iii) The use of advanced technologies is not part of the success criteria of a project. In order to establish a solid ground for giving agent technologies these characteristics, we recognize that accepted methods for industrial development depend on standard representations of artifacts supporting all phases of the software lifecycle. Standard representations are needed by tool developers to provide commercial quality tools that mainstream software engineering departments need for industrial agent systems development.