Jul 08

Distributed Real-time Collaborative Editing System @ Kno.e.sis

Real-time collaborative editing systems such as GoogleDocs or Etherpad are well known and widely used. They allow a group of people to collaborate on a same document from different places at any time. Unfortunately, these systems have scalability limitations regarding the number of users who can collaborate at the same time on a document or the number of concurrent changes they can handle. However, we observe nowadays a change of scale on the numbers of people collaborating on a document: besides small groups, communities of users are now using these tools to coordinate and achieve their writing activities.

In this talk, I will motivate the need for peer-to-peer real-time collaborative editing systems. I will give a brief overview of optimistic replication mechanisms suitable for this kind of systems and give a focus on a conflict-free replicated data type: a sequence whose atoms have an adaptive granularity. Finally, I will present our current prototype: the MUTE collaborative editor — a peer-to-peer real-time web collaborative editor.

Jul 05

Large-scale Trust-based Collaboration @ Kno.e.sis

Distributed collaborative systems such as wikis, version control systems or GoogleDrive allow users to collaborate on a set of shared documents from different places, at any time and from different devices. Existing collaborative systems made available by large service providers such as Google are based on a central authority that stores and has control over user personal data. Moreover, these systems do not scale well in terms of the number of users and the frequency of their modifications. Coast Inria team investigates and designs peer-to-peer collaborative systems that offer a very good scalability and where users share data directly with the other users they trust without relying on a central authority.

In the first part of the talk I’ll present a study on investigating scalability of existing real-time collaborative systems in terms of delays experienced by users, i.e. the time between a modification done by a user is visible to other users. By means of user studies where we simulated delays on small collaborative tasks, we studied the effect of delay on users.  We found out a general effect of delay on performance related to the ability to manage redundancy and errors across the document. We interpret this finding as a compromised ability to maintain awareness of team member activity, and a reversion to independent work.

In the second part of the talk I’ll discuss about a preliminary validation of a trust-based collaboration model where users share data with the people they trust. Trust can be computed based on the collaboration interactions between users. We studied trust game, a money exchange game, widely used in behavioral economics for analysing trust and collaboration between humans. In this game, exchange of money is entirely attributable to the existence of trust between users. We proposed a trust metric that computes trust scores between users based on the amounts they exchanged in the past. This trust metric deals with fluctuating user behavior and can predict future user behavior. By means of user studies we investigated the influence of showing a partner trust score on user behavior during the trust game. We showed that the availability of trust score has the same effect as of user nickname to improve cooperation between users. We conclude that in the case of large scale collaboration where users can change their nicknames and it is difficult to remember who is who from the nicknames, trust scores could be an enhancement over traditional nicknames.

This work has been done in collaboration with Valerie L. Shalin from the Department of Psychology of Wright State University, Kno.e.sis, in the context of the USCOAST associated Inria team.

Jun 28

We are visiting Kno.e.sis @ WSU

In the context of the USCoast Inria associate team, Claudia and Gerald are visiting for a month our colleagues from the Ohio Center of Excellence in Knowledge-enabled Computing (Kno.e.sis) at Wright State University.

wsu

Jun 16

WETICE 2016 Keynote – From group collaboration to large scale social collaboration

Keynote presentation

Jun 16

We were at Inforsid

We attend the Inforsid/RCIS 2016 Conference where Guillaume Rosinosky presented his work at the PhD forum

DSC03389

DSC03482

May 04

PhD Defense Luc André – vendredi 13 mai en salle A008 à 9h30

Préservation des Intentions et Maintien de la Cohérence des Données Répliquées en Temps Réel

Résumé en français :

L’édition collaborative en temps réel permet à plusieurs utilisateurs d’éditer un même document simultanément grâce à des outils informatiques. Les applications d’édition collaborative en temps réel, telles GoogleDocs ou Etherpad, répliquent les données éditées chez chaque utilisateur, pour garantir une édition des données réactive et possible à chaque instant. Les conflits d’édition sont fréquents, et doivent être gérés automatiquement par l’application. L’application doit faire converger toutes les répliques vers un document commun, qui contient toutes les modifications exprimées par tous les utilisateurs.

Les algorithmes actuels fonctionnent de manière satisfaisante pour des types de données simples (des documents linéaires) et des possibilités d’édition minimes (insérer ou supprimer du texte). Lorsque le document est plus complexe (document XML, texte structuré), ou qu’il peut être édité avec un ensemble élargi d’opérations (déplacement de texte, styliser du texte), lors de la résolution de conflits d’édition, les algorithmes échouent à proposer un contenu qui convienne aux utilisateurs. Les intentions des utilisateurs ne sont pas respectées. L’objectif de cette thèse est de proposer des algorithmes d’édition collaborative en temps réel qui respectent mieux les intentions des utilisateurs que les algorithmes actuels.

La première contribution de cette thèse est un algorithme basé sur l’approche des transformées opérationnelles (OT), qui permet d’éditer et de styliser du texte hiérarchisé tout en respectant les intentions de style, de déplacement ou d’insertion de texte ou de paragraphes.

La deuxième contribution est basée sur l’approche des modèles de données répliquées commutatives (CRDT), et propose un algorithme capable de respecter l’intention de mise à jour d’un élément, tout en améliorant les performances globales de l’approche par rapport aux autres CRDTs.

Abstract in English :

Real-time collaborative editors, like GoogleDocs or Etherpad, allow the simultaneous edition of a document by several users. These applications need to replicate the edited document, for the so called real-time purpose of permitting a fast and reactive editing by any user at any time. Editing conflicts frequently occur, and must be automatically handled by the application, in order to provide every users with the same copy of the document, containing every modifications issued.

Most of current real-time collaborative editing algorithms were designed for simple data structures, like linear text, and simple editing ways, like inserting or removing a character only. These algorithms fail to offer an appropriate editing conflict resolution when used with a complex data structure, like XML, or with complex operations, like moving some text or adding some style. Copies are the same but users’ intentions are not preserved. The goal of this thesis is to design new real-time collaborative editing algorithms that ensure a better preservation of users’ intentions.

The first contribution of this thesis is an algorithm based on the Operational Transformation approach (OT). Our contribution is designed to handle rich text document (with stylized text and paragraphs) and to preserve the intentions of a set of high editing level operations (add style, merge paragraphs…).

The second contribution uses the Commutative Replicated Data Types approach (CRDT), and offers an algorithm which preserves the update intention, while improving global performance of the approach when dealing with large blocs of data.

Apr 13

Séminaire Hien Truong/25 avril/10h/A008 Security based on Contextual Co-presence Detection

Séminaire Hien Truong/25 avril/10h/A008

Title: Security based on Contextual Co-presence Detection

Abstract: Although the security research community no longer takes security and usability to be mutually contradictory goals, simultaneously accomplishing security and usability goals continues to be a challenge. We addressed this challenge by investigating a promising approach: exploiting contextual information. We began by addressing a  specific example problem: zero-interaction authentication schemes are essential for usability in certain situations, but are vulnerable to relay attacks. “Passive keyless entry and start” systems such as “Keyless- Go” are intended to increase the usability of car access control systems. They allow the owner of a car to unlock car doors merely by physically approaching the car without having to take the key out to perform any other action, such as pressing a button on the car key. This is an instance of proximity-based “zero interaction authentication” (ZIA). The standard defense against relay attack using techniques for distance-bounding is often impractical. A more realistic approach is to have the parties sense their respective current ambient context along multiple modalities and compare them to see if they can conclude that they are co-present in the same context. We took an experimental approach to answer basic question of “how to determine contextual co-presence?” which requires answering a number of sub-questions. Zero interaction authentication approaches are deployed predominantly based on the verifier detecting the proximity of the user’s personal device, or a security token, by running authentication protocol over a short-range wireless communication channel. We proposed a solution using multiple sensors to collect contextual data to detect co-presence of the prover and verifier. Modern computing devices are equipped with many “sensors” like microphones, wireless networking interfaces, global positioning system (GPS) receivers and so on. A device can extract information from such sensors that are characteristics of context. By having two mutually trusting devices exchange and compare context information, they can determine if they are co-present or not. Although prior works constitute an important step towards addressing the hard problem of resisting relay attacks using off-the-shelf hardware, they leave several important questions unexplored. We did the first work that fairly compares the performance of different sensor modalities in resisting relay attacks against ZIA based on contextual co-presence. We proved the feasibility of using different context (Radio Frequency, physical ambient) to design security solution resisting relay attacks.

Bio

Hien Truong received the Ph.D of Computer Science from INRIA and Universite de Lorraine in France in 2012. Since then she has been working as a postdoctoral researcher at the Department of Computer Science, University of Helsinki in Finland. Her research interests include Information Security, Privacy and Trust for Distributed Systems, and Mobile Security. In current research, she focuses on applying machine learning and data analysis techniques to solve various security problems such as predicting mobile malware, defending against relay attacks in wireless communication channels. She has conducted research activities collaboratively with industrial partners including Intel Collaborative Research Institute for Secure Computing (ICRI-SC) in Finland and NTT Cyber Solutions Laboratories (NTT Labs) in Japan.

Apr 11

The Coast Team at RUE 2016

We did a demo of MUTE at the “Rencontre Université Entreprise” in April 2016 with 3 ChromeBook (TM), a Raspberry Pi (TM) and a Wifi Router

Inria-0237-007Inria-0237-013

 

Inria-0237-006  Inria-0237-008 Inria-0237-012

© Inria / Photo G. Scagnelli

Mar 10

Claudia-Lavinia Ignat à l’honneur dans la Semaine

La Semaine du 10/3/201611_NCY305r

Feb 12

Seminar by Jerome Dinet : Human Factors and Information Security: Impact of Individual, Social and Culture Environment

22/2/2016 – 14h – B013

From identity theft and fraud to corporate hacking attacks, cybersecurity has never been more important for businesses, organizations and governments. But if security and risk can be objectively defined (e.g., by analyzing objective data), risk perception is more important to explain human’s opinions, attitudes and in fine, our behaviors and decision of governments whatever the domain (e.g., the impact of potential health risks from exposure to power-frequency electromagnetic fields, the trust of drivers in autonomous cars, the attitudes towards robots in our daily environment). Risk perception is the subjective judgment that people make about the characteristics and severity of a risk. Risk perception is most commonly used in reference to natural-physical hazards and threats to the environment of health and security. Three major families of theory have been proposed to explain why different people make different estimates of the dangerousness of risks: (1) psychology approaches (heuristics and cognitive biases), (2) anthropology/sociology approaches (cultural theory and impact of organization safety culture) and (3) interdisciplinary approaches (social amplification of risk framework and the role of medias). In other words, the numerous technical advances in information sciences produce more secure environments; but information security cannot be understood or described as solely a technical problem. Human factors must be included because recent research shows that risk perceptions are mostly influenced by the emotional state of the perceiver/user.

Jerome Dinet is Professor of psychology in University of Lorraine (Lab InterPsy, EA4432). His research topics mainly concern human factors involved in human-machine interaction, and the behaviours and cognitive processes during information retrieval of users with specific needs (young children, seniors) and/or with impairments. He was invited professor in New York University (USA), in the National Advanced Institute of Science and Technology of Tsukaba (Japan), and in the University of Uppsala (Sweden). Currently, he is director of a new series entitled “Human-Machine Interaction” for ISTE-Wiley.