Search

Most commonly used collaborative systems are provided by large service providers. While these collaborative services offer very interesting functionalities, they feature certain limitations. Most of the platforms hosting these collaboration services rely on a central authority and placepersonal information in the hands of a single large corporation which is a perceived privacy threat. Users must provide and store their data to vendors of these services and have to trust that they will preserve privacy of their data, but they have little control over the usage of their data after sharing it with other users. Moreover, these systems do not scale well in terms of the number of users and their modifications. Furthermore, user communities cannot deploy these kind of service applications since they generally rely on costly infrastructures rather than allowing sharing infrastructure and administration costs. My research work aims to move away from centralized authority-based collaboration towards a large scale trust-based peer-to-peer collaboration where control over data is given to users who can decide with whom to share their data. The main advantages of peer-to-peer collaborative systems are high scalability, resilience to faults and attacks and a low deployment barrier for new services. My contributions are structured around two axes of research: collaborative data management and trustworthy collaboration.In order to provide efficient data availability, data is typically replicated and users are allowed to concurrently modify replicated data. One of the challenges is to develop optimistic replication algorithms that maintain consistency of the shared data in the face of concurrent modifications. These algorithms have to be reliable, i.e. after the reception of all modifications the data copies have to converge. These algorithms also have to be scalable, i.e. have a complexity that does not depend on the number of users. These algorithms have also to be explainable, i.e. their decision must be comprehensible by users and therefore user intentions must be preserved. In the first part of this manuscript I present my contributions to the design of various optimistic data replication and their evaluation in terms of their complexities but also by means of experimental design with users. I also present my contributions on group awareness specifically on what information should be provided to users to prevent conflicting changes and to understand divergence when conflicts cannot be avoided.In this large scale peer-to-peer collaboration where users cannot remember the interactions with all collaborators a main question is how to choose the trusted collaborators in order to share them the data. A main challenge is how to assess users trust according to their past behaviour during collaboration in order to be able to predict their future behavior. In the second part of this manuscript I present a contract-based collaboration model where contracts are specified by the data owners when they share the data and user trust is assessed according to the observation of adherence to or violation of contracts. I also present a hash chain based authenticators approach for ensuring integrity and authenticity of logs of collaboration. For testing the proposed trust-based collaboration model, I designed a user experiment employing trust game and relying on a computational trust metric according to user exchanges in this game.I finally present my future research directions around secure and trustworthy collaborative data management.

International audience; In distributed collaborative systems, participants maintain a replicated copy of shared documents. They edit their own copy and then share their modifications without any coordination. Copies follow successions of divergence and convergence. Convergence is a liveness property of collaborative systems. Some malicious participants may find an advantage to make the collaboration fail. To that end, they can preclude convergence of the copies. To protect convergence of copies, participants can exploit an authenticated log of modifications. New participants have to retrieve the entire log in order to contribute. Unfortunately, the cost of joining a collaboration increases with the size of this log. Causal Stability allows to prune authenticated logs in a static collaborative group without any malicious participants. In this paper, we first tailor Causal Stability to dynamic groups in presence of malicious participants. We then propose a mechanism to verify the consistency of a pruned log and a mechanism to authenticate a snapshot from a pruned log.

International audience; Distributed collaborative systems allow users to collaborate on a set of shared documents from any place, at any time and from any device. Examples of collaborative systems are wikis, version control systems or GoogleDrive. While 30 years ago when these collaborative systems were firstly developed they were used in scenarios involving only a small set of users such as for the writing of a research article, nowadays we notice a change in the scale from several users to a community of users. The large-scale collaboration is now possible due to advances in mobile and ubiquitous communication that enable users to be continuously connected and to the appropriation of existing tools by the users. However, existing collaborative systems face several challenges including privacy issues as personal user information is placed in the hands of large corporations and users have little control over the usage of their data, performance and coordination issues in the large-scale context. In this talk we will illustrate the evolution of collaborative systems in the past years and describe our vision of trustworthy distributed collaborative systems where communities of users can safely and confidently collaborate without the use of a central authority. We will focus on envisaged solutions for replicated data consistency, security, trust and awareness in this context. As human factor is a key issue in the design of trustworthy distributed collaborative systems, we call for the need of evaluation of these systems with user studies.

The final publication is available at Springer via http://dx.doi.org/10.1007/978-3-540-85257-5_17, Proceedings of the 2nd International Symposium on Intelligent Distributed Computing – IDC 2008, Catania, Italy, 2008, This paper describes a distributed collaborative wiki-based platform that has been designed to facilitate the development of Semantic Web applications. The applications designed using this platform are able to build semantic data through the cooperation of different developers and to exploit that semantic data. The paper shows a practical case study on the application VPOET, and how an application based on Google Gadgets has been designed to test VPOET and let human users exploit the semantic data created. This practical example can be used to show how different Semantic Web technologies can be integrated into a particular Web application, and how the knowledge can be cooperatively improved., This work has been supported by TIN 2005-06885, TIN 2007-64718 and TIN 2007-65989.

International audience; In distributed collaborative systems, replicated objects, shared by users, are subject to concurrency constraints. All methods [4, 13, 18, 15, 16, 19, 22] proposed to serialize concurrent operations and achieve copies convergence of replicated objects are based on the use of Operational Transformations. In this context, giving the user the ability to undo an operation has been recognized as a difficult problem [1, 2, 3, 12, 14, 20, 21]. The few general propositions to solve the problem sometimes compromise copies convergence and/or users’ intention, insofar as the Operational Transformations used are unsuitable for undo. This paper has a twofold objective. Firstly, it aims to highlight two general conditions (named C3 and C4) that need to be satisfied by any transformation adapted to undo. Secondly, it presents a general undo algorithm based on the definition of a generic undo-fitted transformation, which automatically verifies these conditions. The interest of the proposed method is that the undoing of an operation obeys to the same processing as the one used for regular operations in collaborative systems such as [15,19].

International audience; Crowdsourcing is a collaborative business process model, in which tasks are carried out by a crowd. In crowdsourcing systems, there are two types of stakeholders namely, requesters who outsources tasks, and the crowd or contributors, performing those tasks. We consider a stakeholder as an actor or a standalone software component, evolving on a platform and having both mechanisms of interaction with its environment and business skills. A set of stakeholders interacting in a dynamic context for solving a problem, is a distributed collaborative system, and we term it crowsourcing system. In such a system, the role concept is central, because each stakeholder must have a specific framework within which he collaborate. Traditionally, collaborative systems lose flexibility if their design is role-based, because only static role description mechanisms based on intuitive concepts are available. We propose in this paper, an improvement consisting of four things: (1)defining clearly what an outsourceable task or crowd task is, (2)specifying roles clearly and rigorously, while ensuring flexibility for collaboration, (3)providing role switching mechanisms and, (4) providing an abstract basis, for crowdsourcing system design and workflow monitoring and checking mechanisms, for potential activities, dynamically carried out by a system.; Le crowdsourcing est un modèle de processus métier collaboratif, dans lequel les tâches sont externalisées. Dans ces systèmes, on distingue deux types d'intervenants: les demandeurs qui externalisent des t\^{a}ches et la foule ou contributeurs qui effectuent ces tâches. Nous considérons un intervenant à la fois comme un acteur et un composant logiciel autonome, évoluant sur une plateforme, disposant de mécanismes d'interaction avec son environnement et de compétences m\'{e}tiers. Un ensemble d'intervenants interagissant dans un contexte dynamique pour résoudre un problème, est un système collaboratif distribué,désigné par système de crowdsourcing. Dans un tel système, le concept de rôle est central, car chaque intervenant doit disposer d'un cadre spécifique dans lequel il collabore. Traditionnellement, les systèmes collaboratifs perdent en flexibilité si leur conception est rôle-centrée, car seuls les mécanismes statiques de description de rôles, basés sur des concepts intuitifs sont disponibles. Nous proposons dans cet article, une amélioration consistant en quatre choses: (1)définir clairement ce qu'est une tâche externalisable, (2)spécifier les rôles clairement et rigoureusement, tout en assurant la flexibilité de la collaboration, (3) fournir des mécanismes de changement de rôle et, (4) fournir une base abstraite, pour la conception de systèmes de crowdsourcing et les mécanismes de monitoring et de vérification de workflows, pour les activités menées dynamiquement par un système.