fr Thu, 14 Feb 2013 09:54:43 +0100 http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=151 http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=151 26/02/2013 - __/__/____ - CNAM - 33.1.19
La pr?sentation sera faite par Darell Long, professeur ? l'universit? de Santa Cruz et titulaire de la chaire de recherche "Kumar Mallavali" pour les syst?mes de stockage. Ci-joint les deux r?sum?s de pr?sentation pr?vues pour le 26. Abstract #1 RESAR Storage: a System for Two-Failure Tolerant, Self-Adjusting Million Disk Storage Clusters The demand for large-scale storage is greater than ever. The wide availability of broadband networking has made cloud based storage a vibrant and growing market. Additionally, as we explore exascale high performance computing (HPC) systems with exabytes of data, power considerations become a significant factor. Most existing systems rely on replication to protect user data, maintaining as many as six copies. This high overhead leads to an unnecessary costs in equipment, maintenance and energy. While storage appliances using era- sure coding schemes are available, their long rebuild times and lack of continuity of service during rebuild make them unsuitable as building blocks for large scale storage systems. We present RESAR (Robust, Efficient, Scalable, Autonomous Reliable) storage, a reliable distributed storage volume provider that scales to millions of drives. We implemented our system and tested it on a large-scale emulation platform called Megatux. Our results show that RESAR is capable of scaling to millions of drives, and it?s rebuild performance benefits from this scale by distributing the recovery across many disks. In our emulations, the work of rebuilding a one terabyte hard drive was distributed across 400 disks and completed in less than four minutes with no interruption of service. With an annual durability of 99.999999% and a storage overhead cost of 20%, RESAR has great promise for both exascale HPC and cloud storage. Joint work with my Ph.D. student Igancio Coder?, Thomas Kroeger of Sandia National Laboratory and Thomas Schwarz of Universidad Cat?lica del Uruguay. Abstract #2 Horus: Fine-Grained Encryption-Based Security for Large-Scale Storage With the growing use of large-scale distributed systems, the likelihood that at least one node is compromised is increasing. Large-scale systems that process sensitive data such as geographic data with defense implications, drug modeling, nuclear explosion modeling, and private genomic data would benefit greatly from strong computing (HPC), cloud, or secure content delivery network (SCDN) systems that handle such data still store them unencrypted or use simple encryption schemes, relying heavily on physical isolation to ensure confidentiality, providing little protection against compromised computers or malicious insiders. Moreover, current encryption solutions cannot efficiently provide fine-grained encryption for large datasets. Our approach, Horus, encrypts large datasets using keyed hash trees (KHTs) to generate different keys for each region of the dataset, providing fine-grained security: the key for one region cannot be used to access another region. Horus also reduces key management and distribution overhead while providing end-to-end data encryption and reducing the need to trust system operators or cloud service providers. Horus requires little modification to existing systems and user applications. Performance evaluation shows that our prototype's key distribution is highly scalable and robust: a single key server can provide 140,000 keys per second, theoretically enough to sustain more than 100 GB/s I/O throughput, and multiple key servers can efficiently operate in parallel to support load balancing and reliability. Horus will appear in FAST 2013. It is joint work with Yan Li, Nakul Sanjay Dhotre, Yasuhiro Ohara,Thomas Kroeger, and Ethan Miller Biography Darrell D. E. Long is a Fellow of the Institute of Electrical and Electronics Engineers and of the American Association for the Advancement of Science He is the Director of the Storage Systems Research Center at the University of California, Santa Cruz where he is Professor of Computer Science and holds the Kumar Malavalli Endowed Chair. He received his B.S. degree in Computer Science from San Diego State University, and his M.S. and Ph.D. from the University of California, San Diego. His current research interests in the storage systems area include high performance storage systems, archival storage systems and exascale file systems. His research also includes computer system reliability, video-on-demand, applied machine learning and computer security. He served as the Vice Chair and then Chair of the University of California Committee on Research Policy. He has served on the University of California President?s Council on the National Laboratories, and on the Science & Technology, National Security and Intelligence committees. He currently serves on the Science & Technology committee for both Los Alamos and Lawrence Livermore National Laboratories. He served on the National Research Council Standing Committee on Technology Insight-Gauge, Evaluate and Review. Thu, 14 Feb 2013 09:54:43 +0100 Vertigo Event http://cedric.cnam.fr/index.php/labo/membre/view?id=1762 http://cedric.cnam.fr/index.php/labo/membre/view?id=1762 a Tue, 29 Jan 2013 11:48:51 +0100 Vertigo Job http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=150 http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=150 29/01/2013 - __/__/____ - CNAM - 33.1.18
La pr?sentation sera faite par Hicham RANDRIANARIVO, doctorant du CNAM ? l'ONERA sous la direction de Michel Crucianu, et encadr? par Bertrand Le Saux et Marin Ferecatu. En voici le r?sum? : "Nous nous int?ressons au probl?me de la reconnaissance et la localisation d'objets dans des images. Aujourd'hui les m?thodes qui rencontrent le plus de succ?s en d?tection d'objet sont celles bas?es sur une repr?sentation de l'objet par mod?le ? parties d?formables. Ces m?thodes ont montr? leurs succ?s sur des bases d'image tel que Pascal VOC ou Inria Person, nous souhaiterions dans un premier temps adapter ces m?thodes pour la d?tection de structures artificielles dans des images satellitaire puis dans un second temps am?liorer les performances de ces m?thodes en utilisant une nouvelle approche pour l'apprentissage du mod?le: le co-apprentissage." Thu, 24 Jan 2013 16:30:48 +0100 Vertigo Event http://cedric.cnam.fr/index.php/publis/article/view?id=2682 http://cedric.cnam.fr/index.php/publis/article/view?id=2682 Nous présentons RoSeS (Syndication Web Réellement Simple, Libre et Efficace), un sys- tème pour l?interrogation et l?agrégation de flux RSS. RoSeS est fondé sur un langage déclaratif pour la définition de requêtes continues sur des collections de flux RSS. Les utilisateurs créent des flux personnalisés en composant des requêtes de filtrage et d?agrégation. La publication d?une requête sous forme d?une vue et sa réutilisation dans la définition d?autres requêtes re- flète naturellement le principe de publication-souscription appliqué dans la syndication de flux RSS. Nos contributions principales sont un langage déclaratif d?agrégation de flux RSS, une algèbre extensible favorisant la création de plans d?exécution efficaces, une stratégie d?optimi- sation multi-requêtes et un prototype complet intégrant un moteur de requêtes continues. Thu, 22 Nov 2012 11:18:06 +0100 Vertigo Paper http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=147 http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=147 14/11/2012 - 15/11/2012 - Amphi Y - Z, Cnam
  Wed, 14 Nov 2012 15:55:14 +0100 Vertigo Event http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=146 http://cedric.cnam.fr/index.php/default/cms_evenement/view?id=146 20/11/2012 - 20/11/2012 - 33.1.19
"Indexes Analysis for Matching Subscriptions in RSS feeds" Mon, 05 Nov 2012 16:00:04 +0100 Vertigo Event http://cedric.cnam.fr/index.php/publis/article/view?id=2677 http://cedric.cnam.fr/index.php/publis/article/view?id=2677 toto Mon, 05 Nov 2012 15:56:38 +0100 Vertigo Paper http://cedric.cnam.fr/index.php/publis/article/view?id=2676 http://cedric.cnam.fr/index.php/publis/article/view?id=2676 vezrverzvze Mon, 05 Nov 2012 15:52:26 +0100 Vertigo Paper http://cedric.cnam.fr/index.php/labo/membre/view?id=1721 http://cedric.cnam.fr/index.php/labo/membre/view?id=1721 a Fri, 05 Oct 2012 12:58:25 +0200 Vertigo Job http://cedric.cnam.fr/index.php/labo/membre/view?id=1720 http://cedric.cnam.fr/index.php/labo/membre/view?id=1720 a Fri, 05 Oct 2012 12:56:59 +0200 Vertigo Job http://cedric.cnam.fr/index.php/labo/membre/view?id=1700 http://cedric.cnam.fr/index.php/labo/membre/view?id=1700 a Wed, 26 Sep 2012 09:11:13 +0200 Vertigo Job http://cedric.cnam.fr/index.php/publis/article/view?id=2615 http://cedric.cnam.fr/index.php/publis/article/view?id=2615 The demonstration is devoted to the desktop-level interactions offered by Cador, a content-based document management system currently under development. Cador provides a rule-based language to query and manipulate large collections of documents distributed in repositories. The language is able to define the content of Virtual File Systems (VFS) as views over the document collections. This feature allows users to combine their familiar interface and desktop-based softwares with the powerful search and transformation tools provided by the underlying system. The demonstration shows how VFS views can be created on-demand to present a desktop-based virtual document organization and how standard desktop interactions can be captured and interpreted in terms of document management operations: creation, updates, annotation, derivation of new content thanks to transformation rules, sharing between users, etc. The example application is the management of a large bibliographic database: users can, with a few clicks, organize their bibliographic references, import new references, share them with a group of co-authors and automatically maintain a ready-to-use Bibtex file. Thu, 12 Jul 2012 15:46:57 +0200 Vertigo Paper http://cedric.cnam.fr/index.php/publis/article/view?id=2614 http://cedric.cnam.fr/index.php/publis/article/view?id=2614 The explosion of published information on the Web leads to the emergence of a Web syndication paradigm, which transforms the passive reader into an active information col- lector. Information consumers subscribe to RSS/Atom feeds and are notified whenever a piece of news (item) is published. The success of this Web syndication now offered on Web sites, blogs, and social media, however raises scalability issues. There is a vital need for efficient real-time filtering methods across feeds, to allow users to follow effectively personally interesting information. We investigate in this paper three indexing techniques for users' subscriptions based on inverted lists or on an ordered trie. We present analytical models for memory requirements and matching time and we conduct a thorough experimental evaluation to exhibit the impact of critical workload parameters on these structures. Thu, 12 Jul 2012 15:41:54 +0200 Vertigo Paper http://cedric.cnam.fr/index.php/default/recrutement/description?id=9 http://cedric.cnam.fr/index.php/default/recrutement/description?id=9 a Wed, 11 Jul 2012 21:05:19 +0200 Vertigo Job http://cedric.cnam.fr/index.php/publis/article/view?id=2595 http://cedric.cnam.fr/index.php/publis/article/view?id=2595 Un système d'information (SI) est un ensemble organisé de ressources incluant non seulement des ressources numériques (logiciels, machines, données) mais aussi des ressources humaines. Dans notre vision l'humain n'est pas un "simple utilisateur", il est aussi un compo- sant du système, porteur de connaissances. Cette vision nous conduit à introduire le concept de Système d'Information et de Connaissances (SICO), et à envisager son évaluation. Pour cela, nous choisissons un cadre d'évaluation du système d'information de la littérature, reposant sur une vision du SI proche de celle du SICO, que nous étendons afin d'inclure un nouveau point de vue de l'évaluation : celui de la prise en compte des connaissances. Nous proposons ensuite un modèle pour l'évaluation de ce point de vue. Ce modèle permet d'apprécier le niveau de maturité de la prise en compte des connaissances au sein d'un SI. Wed, 20 Jun 2012 20:53:58 +0200 Vertigo Paper http://cedric.cnam.fr/index.php/publis/article/view?id=2593 http://cedric.cnam.fr/index.php/publis/article/view?id=2593 Un grand nombre de définitions du système d'information (SI) ont été proposées dans la littérature. Il est maintenant bien accepté par les communautés manipulant ce concept que le système d'information combine un système d'information numérique et des activités humaines afin de servir de support aux branches opérationnelles, à la gestion et à la prise de décision dans l'entreprise. Assurer la qualité du SI est donc crucial. Cependant, évaluer cette qualité est une tâche multidimensionnelle complexe, concernant différentes ressources (processus, données, artéfacts, interfaces, ...), considérant différentes caractéristiques (utilisabilité, fonctionnalités, ...) et mettant en oeuvre différentes méthodes pour effectuer l'évaluation. Après plusieurs années d'expérience dans l'évaluation des SI, nous résumons dans ce papier les leçons que nous avons tirées des évaluations que nous avons menées, et exhibons les limites de l'évaluation d'un SI. Nous finissons le papier par une réflexion sur l'avenir de l'évaluation d'un SI. Wed, 20 Jun 2012 15:29:15 +0200 Vertigo Paper http://cedric.cnam.fr/index.php/default/recrutement/description?id=8 http://cedric.cnam.fr/index.php/default/recrutement/description?id=8 a Wed, 16 May 2012 09:43:23 +0200 Vertigo Job http://cedric.cnam.fr/index.php/publis/article/view?id=2539 http://cedric.cnam.fr/index.php/publis/article/view?id=2539 This report is focused on the study of methods for image retrieval in collection of heterogeneous contents. The spatial relationships between entities in an image allow to create the global description of the image that we call the image context. Taking into account the contextual spatial relationships in the similarity search of images can allow improving the retrieval quality by limiting false alarms. We defined the context of image as the presence of entity categories and their spatial relationships in the image. By studying statistically the relationships between different entity categories on LabelMe, a symbolic images databases of heterogeneous content, we create a cartography of their spatial re- lationships that can be integrated in a graph-based model of the contextual relationships, the principal contribution of this report. This graph describes the general knowledge of every entity categories. Spatial reasoning on this knowledge graph can help improving tasks of image proces- sing such as detection and localization of an entity category by using the presence of another one. Further, this model can be applied to represent the context of an image. The similarity search ba- sed on context can be achieved by comparing the graphs, then, contextual similarity between two images is evaluated by the similarity between their graphs. This work was evaluated on the symbo- lic image database of LabelMe. The experiments showed its relevance for image retrieval by context. Fri, 11 May 2012 10:39:37 +0200 Vertigo Paper http://cedric.cnam.fr/index.php/publis/article/view?id=2495 http://cedric.cnam.fr/index.php/publis/article/view?id=2495 In the context of computer-assisted plant identification we are facing challenging information retrieval problems because of the very high within-class variability and of the limited number of training examples. To address these problems, we suggest a new interactive learning approach that combines similarity-based retrieval and re-ranking by SVM using local feature distributions. This approach leads to improved sample selection, allowing to obtain better results. Tue, 03 Apr 2012 14:55:13 +0200 Vertigo Paper http://cedric.cnam.fr/index.php/publis/article/view?id=2492 http://cedric.cnam.fr/index.php/publis/article/view?id=2492 test Mon, 02 Apr 2012 13:44:40 +0200 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2488 http://cedric.dev/index.php/publis/article/view?id=2488 test Fri, 23 Mar 2012 15:09:07 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2487 http://cedric.dev/index.php/publis/article/view?id=2487 test Fri, 23 Mar 2012 15:07:42 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2486 http://cedric.dev/index.php/publis/article/view?id=2486 test Fri, 23 Mar 2012 15:03:08 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2485 http://cedric.dev/index.php/publis/article/view?id=2485 test Fri, 23 Mar 2012 15:01:59 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2484 http://cedric.dev/index.php/publis/article/view?id=2484 test Fri, 23 Mar 2012 15:00:55 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2483 http://cedric.dev/index.php/publis/article/view?id=2483 test Fri, 23 Mar 2012 15:00:02 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2482 http://cedric.dev/index.php/publis/article/view?id=2482 test Fri, 23 Mar 2012 14:59:03 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2481 http://cedric.dev/index.php/publis/article/view?id=2481 test Fri, 23 Mar 2012 14:58:02 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2480 http://cedric.dev/index.php/publis/article/view?id=2480 test Fri, 23 Mar 2012 14:57:10 +0100 Vertigo Paper http://cedric.dev/index.php/publis/article/view?id=2479 http://cedric.dev/index.php/publis/article/view?id=2479 test Fri, 23 Mar 2012 14:56:14 +0100 Vertigo Paper