Retour au site du LAAS-CNRS

Laboratoire d’analyse et d’architecture des systèmes
Choisir la langue : FR | EN

485documents trouvés

17302
21/11/2017

Cross fertilization between wireless testbeds and NS-3 simulation models

G.KREMER, P.OWEZARSKI, P.BERTHOU

SARA

Manifestation avec acte : ACM International Conference on Modeling, Analysis and Simulation of Wireless and Mobile Systems ( MSWIM ) 2017 du 21 novembre au 25 novembre 2017, Miami (USA), Novembre 2017, 9p. , N° 17302

Lien : https://hal.laas.fr/hal-01575822v1

Diffusable

Plus d'informations

Abstract

Network simulators are ooen used for their simplicity and cost regarding wireless networks. However, their realism is ooen criticized and their results challenged. e main concern comes from the modeling of the PHY and MAC layers. To assess the performances of these simulators and their models, the results of simulations are ooen compared with experimental results. However, the comparison methodolo-gies used in these studies may introduce biases. is work focuses on accurately discovering and analyzing the reasons for the calibration problems or implementation bugs in sim-ulators and experimental devices. For this purpose, we leverage the famous Root Cause Analysis (RCA) technique for comparing traces issued from diierent simulations and real experiments, that includes the study of the root causes of dis-similarities. roughout the paper, our RCA-based method has been applied to detect and analyze a performance anomaly between NS-3 simulation and our lab wireless testbed when transmiiing data over a WIFI 802.11 link. It especially details how low level traac traces have been generated in both environments for similar scenarios, and how they can accurately be compared and their diierences analyzed. ACM Reference format:

140833
17310
21/11/2017

A generic and configurable topology discovery service for software defined wireless multi-hop network

L.CHEN, S.ABDELLATIF, P.BERTHOU, B.NOUGNANKE, T.GAYRAUD

SARA

Manifestation avec acte : ACM International Symposium on Mobility Management and Wireless Access ( MOBIWAC ) 2017 du 21 novembre au 25 novembre 2017, Miami (USA), Novembre 2017, 4p. , N° 17310

Lien : https://hal.laas.fr/hal-01583905

Diffusable

Plus d'informations

Abstract

We present a topology discovery service for software-defined wireless multi-hop networks, which is capable of capturing rich information about the network topology, such as link quality, interference and node characteristics etc., to effectively support the requirements of higher-layer SDN applications such as routing, channel allocation and transmission power control with enhanced flexibility. We analyse the general topology representation required by such SDN applications and specify the procedure and mechanism required at the controller and nodes to maintain this representation. A proof-of-concept implementation of our service on an SDN enabled multi-channel multi-interface wireless multi-hop network testbed shows the feasibility of our proposal.

140915
17388
23/10/2017

Weighted localization in mobile wireless networks

I.MAHJRI, S.MEDJIAH, A.DHRAIEF, K.DRIRA, A.BELGHITH

SARA, HANA Research Group

Rapport LAAS N°17388, Octobre 2017, 6p.

Lien : https://hal.laas.fr/hal-01619274

Diffusable

Plus d'informations

Abstract

Localization of wireless devices is a crucial requirement for many emerging applications such as environmental monitoring, intelligent transportation, home automation, health-care monitoring and social networking. In this letter, we propose AWL a new Aggregate Weighted Localization algorithm for mobile wireless networks. The proposed algorithm is distributed and requires low computational and communication overheads enabling its use in resource-limited devices.

141324
17364
13/10/2017

Autonomic dynamically reconfigurable architectures for future communication systems

K.DRIRA

SARA

Rapport LAAS N°17364, Octobre 2017, 4p.

Lien : https://hal.laas.fr/hal-01592617

Diffusable

Plus d'informations

Abstract

Our objectives involve the development of correct by design models and operational solutions to discover, compose and manage, by automated procedures, the properties of adaptability essential for autonomous reconfiguration. These properties are critical for highly dynamic complex systems such as massive data collection, transport and processing systems , and associated IoT applications. This research direction is of interest for many applications that, on the one hand, have strong reconfiguration requirements to manage mobility, ubiquity, adaptation and self-healing, which are essential for autonomy, and that, on the other hand, are subject to non-functional constraints of quality of service, robustness and reliability. We propose, here, to develop multi-level reconfiguration solutions (VM clusters, application containers, message partitions), in order to manage simultaneously and consistently the adaptation in the different architectural levels. This is necessary for service continuity and end-to-end quality we implemented by coordinating the adaptation actions to satisfy the evolving of the context.

141215
17350
29/09/2017

Distributed localization and conflict detection in mobile wireless networks

I.MAHJRI

SARA

Doctorat : INSA de Toulouse, Septembre 2017, 150p., Président: L.A.GRIECO, Rapporteurs: M.BENJEMAA, F.KRIEF, Examinateurs: A.DHRAIEF, I.JEMILI, Directeurs de thèse: K.DRIRA, S.MEDJIAH , N° 17350

Lien : https://hal.laas.fr/tel-01615478

Diffusable

Plus d'informations

Résumé

Un réseau sans fil mobile est une collection de noeuds mobiles connectés via des liens sans fil. Le noeud mobile peut être une personne portant un terminal mobile, un animal avec une étiquette électronique, un véhicule ou tout autre appareil mobile doté de capacités de communication sans fil. Cette thèse traite deux importants problèmes dans les réseaux sans fil mobiles : la localisation et la détection des conflits. La localisation est l’estimation des positions absolues ou relatives des noeuds mobiles. La détection des conflits est la prédiction des potentiels futurs conflits entre les noeuds mobiles. Un conflit est une situation dans laquelle deux ou plusieurs noeuds mobiles se trouvent à une distance inférieure à une distance minimale requise. La détection des conflits est particulièrement nécessaire dans les réseaux mobiles où les noeuds risquent de trop se rapprocher les uns des autres et se heurter. La première contribution de cette thèse est un nouvel algorithme de localisation pour les réseaux sans fil mobiles. L’algorithme proposé est distribué et nécessite un faible coût de calcul et de communication permettant son utilisation par des noeuds à ressources limitées. En particulier, un petit ensemble de noeuds avec des positions connues, appelés des noeuds ancres, diffusent périodiquement leurs coordonnées. Chaque noeud mobile avec une position inconnue collecte les coordonnées des noeuds ancres, les attribue des poids en fonction de différentes métriques (par exemple, la qualité de la liaison sans fil, le temps de réception) et enfin estime sa position en tant que la moyenne pondérée de toutes les coordonnées collectées. La deuxième contribution de cette thèse consiste en un algorithme de détection de conflit nommé SLIDE. SLIDE est distribué et à faible complexité: chaque noeuds mobile utilise des simples inégalités mathématiques afin de prédire avec précision et en temps opportun les futurs conflits avec les noeuds voisins. En outre, SLIDE abandonne l’hypothèse restrictive des positions et vitesses précises ainsi que des communications non perturbées afin de garantir son efficacité dans les applications du monde réel. La troisième et dernière contribution de cette thèse est un modèle stochastique qui évalue les risques des conflits dans un réseau sans fil mobile où les noeuds mobiles se déplacent dans le même espace partagé. Contrairement à la plupart des modèles stochastiques existants, le modèle proposé est simple, générique et ne nécessite que deux paramètres d’entrée: le nombre de noeuds mobiles et le paramètre λ caractérisant le temps d’inter-contact entre une paire de noeuds mobiles. Le paramètre λ peut être difficile à estimer expérimentalement. Nous proposons donc une expression générique pour λ et puis nous la spécifions pour deux modèles de mobilité couramment utilisés. Toutes les contributions de cette thèse sont validées en utilisant des simulations basées sur le simulateur d’événements discrets OMNeT ++.

Abstract

A mobile wireless network is a collection of mobile nodes connected via wireless links. The mobile node can be a person carrying a mobile terminal, an animal with an embedded sensor (electronic tag), a vehicle or any other mobile device with wireless communication capabilities. This thesis deals with two fundamental issues in mobile wireless networks: localization and conflict detection. Localization is the estimation of the absolute or relative positions of the mobile nodes. Conflict detection is the prediction of potential future conflicts between the mobile nodes. A conflict is a situation in which two or more mobile nodes are within an unsafe distance from one another. Conflict detection is a crucial requirement for mobile networks where the nodes can get too close to each other and collide. The first contribution of this thesis is a new weighted localization algorithm for mobile wireless networks. The proposed algorithm is distributed and requires low computational and communication overheads enabling its use in resource-limited nodes. In particular, a small set of nodes with known positions, called beacon nodes, are periodically broadcasting their coordinates. A mobile node with an unknown position collects the beacon nodes coordinates, attributes them weights based on different metrics (e.g, link quality, reception time) and finally estimates its position as the weighted average of all the collected coordinates. The second contribution of this thesis consists in a straight line conflict detection algorithm called SLIDE. SLIDE is distributed and lightweight: each mobile node uses simple mathematical inequalities in order to accurately and timely predict future conflicts with the surrounding nodes. Furthermore, SLIDE drops the restrictive assumption of perfect sensing capabilities and perturbation-free environment in order to guarantee its efficiency in real world applications. The third and last contribution of this thesis is a stochastic model that assesses the conflicts risks in a mobile wireless network where the mobile nodes are moving in the same shared space. Unlike most of the existing stochastic models, the proposed model is simple, generic and requires only two input parameters: the number of mobile nodes and the parameter λ characterizing the inter-contact time between a pair of mobile nodes. The parameter λ may be difficult to estimate experimentally. We therefore provide a generic explicit expression for λ and then specify this generic expression for two commonly used mobility models. All the contributions of this thesis are validated through extensive simulations based on the discrete-event simulator OMNeT++.

Mots-Clés / Keywords
Mobile wireless networks; Distributed localization; Distributed conflict detection; Réseaux sans fil mobiles; Localisation distribuée; Détection distribuée des conflits;

141137
17269
25/09/2017

A dynamic resource defragmentation scheme for virtualized SDN-enabled substrate networks

A.SIMO TEGUEU, S.ABDELLATIF, T.VILLEMUR, P.BERTHOU

SARA

Manifestation avec acte : IEEE International Conference on Cloud Networking ( CLOUDNET ) 2017 du 25 septembre au 27 septembre 2017, Prague (République Tchèque), Septembre 2017, 6p. , N° 17269

Lien : https://hal.laas.fr/hal-01570578

Diffusable

Plus d'informations

Abstract

Virtual network embedding (VNE) was subject to extensive research which lead to the emergence of a large number of efficient VNE algorithms. When virtual networks (VNs) arrive and depart over time, the substrate network can easily drift into an inefficient configuration, where resources are increasingly fragmented causing a VN request rejection although cumulatively, there are enough available resources. The ability to reallocate running VNs clearly leads to a better resource utilization. In this paper, we propose " Garbage Collector " (GC), a novel network control program for dynamic and online resource management in virtualized SDN-enabled substrates. GC efficiently addresses the fragmentation problem by performing selective migration. Simulations show that GC clearly improves acceptance ratio of VNE algorithms. They also reveal that, it outperforms some existing works from the literature by increasing the VN acceptance ratio by more than 10%.

140649
17316
22/09/2017

Companion Proceedings 11th European Conference on Software Architecture (ECSA 2017)

T.BURES, R.DE LEMOS, K.DRIRA, L.DUCHIEN, R.MIRANDOLA, C.TRUBIANI, J.PEREZ BENEDI

Czech Academy, University of Kent, SARA, LIFL, UPM, GSSI, UPV

Ouvrage (éditeur) : ECSA '17: Proceedings of the 11th European Conference on Software Architecture: Companion Proceedings, Septembre 2017, 2p. , N° 17316

Lien : https://hal.laas.fr/hal-01583947

Diffusable

Plus d'informations

Abstract

The European Conference on Software Architecture (ECSA), held this year on 11 - 15 September 2017 at the University of Kent at Canterbury, United Kingdom, is a premier European software architecture conference, providing researchers, practitioners, and educators with a platform to present and discuss the most recent, innovative and significant findings and experiences in the field of software architecture research and practice. In addition to the main track, it features various events and tracks including the ECSA Track on Women in Software Architecture, Doctoral symposium, Tool demos, Poster session, and four workshops. All these are done with the aim to allow exploring new trends and supporting researchers at their early stage of career.

140937
17315
22/09/2017

Feature model as a design-pattern-based service contract for the service provider in the service oriented architecture

A.KAMOUN, M.HADJ KACEM, A.HADJ KACEM, K.DRIRA

ReDCAD Laboratory, SARA

Rapport LAAS N°17315, Septembre 2017, 26p.

Lien : https://hal.laas.fr/hal-01580264

Diffusable

Plus d'informations

Abstract

In Service Oriented Architecture (SOA), many feature modeling approaches of Service Provider (SP) have been proposed, notably: the two widely used service contracts WSDL and WADL. By studying these approaches , we found that they suffer from several problems, notably: they only work for specific communication technologies (e.g., SOAP or REST) and they do not explicitly model SOA Design Pattern (DPs) and their compounds. One major benefit of using a DP or a compound DP is to develop SPs with proven design solutions. In this paper, in order to overcome these problems, we propose an approach that integrates Software Product Line (SPL) techniques in the development of SPs. Essentially, we propose a Feature Model (FM), which is the defacto standard for variability modeling in SPL, for the feature modeling of SP. This FM, named F M SP , is designed as a DP-based service contract for SP that models different features including 16 SOA DPs and their compounds that are related to the service messaging category. Its objective to enable developers to generate fully functional, valid, DP-based and highly customized SPs for different communication technologies. Through a practical case study and a developed tool, we validate our F M SP and demonstrate that it reduces the development costs (effort and time) of SPs.

140935
17360
21/09/2017

Gestion autonomique de la QoS au niveau middleware dans l’IoT

Y.BANOUAR

SARA

Doctorat : Université de Toulouse III - Paul Sabatier, 21 Septembre 2017, 175p., Président: T.GAYRAUD, Rapporteurs: G.DIAZ, D.DONSEZ, Examinateurs: O.FOURMAUX, T.MONTEIL, Directeurs de thèse: C.CHASSOT , N° 17360

Lien : https://hal.laas.fr/tel-01624249

Diffusable

Plus d'informations

Abstract

The Internet is experiencing a drastic expansion again. In addition to conventional terminals, it now allows to interconnect all kinds of connected objects allowing the capture of events from the considered environment, but also the remote control of this environment. Billions of these objects are thus led in 2020 to contribute to the advent of the Internet of Things (IoT). This paradigm, which extends the Machine-to-Machine (M2M) concept, paves the way for new uses such as home automation, remote monitoring, or even the factories of the future. Several architectures have been proposed to structure the IoT. Their foundation is based on a vision in four levels: (1) Equipment level, which includes the IoT equipment, (2) Network level containing the various technologies for data exchanges, (3) Middleware level, which offers applications an abstraction layer for underlying levels, and finally, (4) Application level, which consists of the set of applications contributing, via their interactions with the connected objects, to the realization of a business activity. Our work is positioned at Middleware level and is based on the architecture defined in the SmartM2M and then oneM2M standards. Several challenges have to be (re)considered in this context. We are mainly interested in the Quality of Service (QoS) issue expressed by some business applications. Proposals addressing this issue essentially target the Network level. For the Middleware level, standards focus mainly on the proposal of architectures and functional services. The non-functional requirements, typically QoS, are little or not considered. Meanwhile, proprietary solutions do not consider the dynamic evolution of the context and requirements. In response to these limitations, we propose a dynamic management approach, i.e. during the execution of the system, and autonomous, i.e. without human intervention. The proposed management, guided by models, focuses on behavioural and structural reconfiguration actions related to application traffic and/or Middleware resources. The first contribution of this thesis deals with the specification, design, implementation and evaluation of QoS management mechanisms. The second contribution is the specification and design of a software architecture for the self-adaptive QoS management at the Middleware level according to one of the SmartM2M and oneM2M standards. This architecture, called IoT-Q, is based on the application of hierarchical policies in the development of adaptation actions, as well as on the paradigm of Autonomic Computing. The third contribution proposes and validates an analytical model, based on the queuing theory, for the OM2M Middleware (open source implementation of the SmartM2M standard). This model allows to estimate the performance of the involved Middleware entities. The application of this model to the monitoring phase is studied under two approaches, reactive and proactive, coupled with Complex Event Processing (CEP) techniques for the first approach, and a ARMA (Auto -Regressive Moving Average) prediction model for the second one. The fourth contribution concerns the planning of resources-oriented mechanisms that we propose to guide following an approach coupling the analytical model with a model of the system based on graphs. This model allows to represent and consider the possible configurations of the involved entities and their interactions. The rules of graphs matching and rewriting that ensure the conformity of the system with the architectural style are also produced. As a fifth contribution, a case study dealing with a crisis situation in an urban transport environment is also proposed to illustrate the application of the proposed models based on graphs.

Résumé

L’Internet connaît à nouveau une expansion drastique. En plus des terminaux classiques, il permet aujourd’hui d'interconnecter toute sorte d’objets connectés permettant la capture d’événements depuis l’environnement considéré, mais également le contrôle à distance de cet environnement. Plusieurs milliards de ces objets sont ainsi amenés à l’horizon 2020 à contribuer à l'avènement de l’Internet des Objets (IoT). Ce paradigme, qui étend le concept de Machine-to-Machine (M2M), ouvre la voie à de nouveaux usages tels que la domotique, la télésurveillance, ou encore les usines du futur. Plusieurs architectures ont été proposées pour structurer l’IoT. Leur fondement est basé sur une vision en quatre niveaux : le niveau Équipement, qui comporte les objets connectés, le niveau Réseau contenant les différentes technologies nécessaires aux échanges, le niveau Intergiciel (ou Middleware) qui offre aux applications une couche d’abstraction des niveaux sous-jacents, et enfin le niveau Application qui consiste en l’ensemble des applications concourant, via leurs interactions avec les objets connectés, à la réalisation d’une activité métier. Nos travaux se positionnent au niveau Middleware sur la base de l’architecture définie dans le cadre des standards SmartM2M puis oneM2M. Plusieurs problématiques sont amenées à être (re)-posées dans ce contexte. Nous nous intéressons essentiellement à celle de la qualité de service (QoS - Quality of Service) exprimée par certaines applications métiers. Les solutions proposées en réponse à cette problématique concernent principalement le niveau Réseau. Au niveau Middleware, les standards se focalisent essentiellement sur la proposition d’architectures et de services fonctionnels. Les besoins non fonctionnels, typiquement orientés QoS, ne sont que peu ou pas considérés. Parallèlement, les solutions propriétaires ne considèrent pas l’évolution dynamique du contexte et des besoins. Face à ces limites, nous proposons une approche de gestion dynamique, i.e. durant l’exécution du système, et autonome induisant un minimum d’intervention humaine. La gestion proposée, guidée par des modèles, porte sur des actions de reconfiguration comportementales et structurelles touchant au trafic applicatif et/ou sur les ressources de niveau Middleware. La première contribution de cette thèse porte sur la spécification, la conception, l’implémentation et l’évaluation de mécanismes de gestion de la QoS. La deuxième contribution consiste en la spécification et la conception d’une architecture logicielle pour la gestion auto-adaptative de la QoS au niveau Middleware suivant l’un ou l’autre des standards SmartM2M et oneM2M. Cette architecture, intitulé IoT-Q, repose sur l’application de politiques hiérarchiques dans l'élaboration des actions d’adaptation, ainsi que sur le paradigme de l’Autonomic Computing. La troisième contribution consiste en la proposition et la validation d’un modèle basé sur la théorie des files d’attente du Middleware OM2M (implémentation open source du standard SmartM2M), permettant d’estimer les performances des entités Middleware impliquées. Son application à la phase de monitoring est étudiée sous deux approches, réactive et proactive, en couplage avec des techniques de CEP (Complex Event Processing) pour la première approche, et d’un modèle de prédiction ARMA (Auto-Regressive Moving Average) pour la deuxième. La quatrième proposition porte sur la planification des mécanismes orientés ressources, que nous proposons de guider par une approche couplant le modèle analytique avec un modèle du système à base de graphes. Ce modèle permet de représenter et de prendre en compte les configurations possibles des entités impliquées et de leurs interactions. Des règles d’appariement et de réécriture de graphes permettant d’assurer la conformité du système avec le style architectural sont également produites. En cinquième contribution, un cas d’étude portant sur une situation de crise dans un environnement de transport urbain est élaboré pour illustrer l’application de ces derniers modèles.

Mots-Clés / Keywords
IoT; Middleware; QoS; Scalabilité; Autonomic computing; Théorie de files d'attente; Théorie des graphes; Réécriture de graphe; Modèles de prédiction; Cloud computing; Queueing theory; Graphs theory; Graph rewriting; Prediction models;

141181
17297
18/09/2017

Integrating short history for improving clustering based network traffic anomaly detection

J.DROMARD, P.OWEZARSKI

SARA

Manifestation avec acte : International Workshop on Autonomic Systems for Big Data Analytics ( ASBDA ) 2017 du 18 septembre au 22 septembre 2017, Tucson (USA), Septembre 2017, 8p. , N° 17297

Lien : https://hal.archives-ouvertes.fr/hal-01576752

Diffusable

Plus d'informations

Abstract

Traffic anomaly detection is of premier importance for network administrators as anomalies have a dramatic impact on network performances, and QoS perceived by users. It is, however, a very time consuming and costly task that often requires decision from network and security experts. For making anomaly detection autonomous, many research works started investigating the use of unsupervised machine learning techniques, and in most cases traffic clustering. Identifying the clusters corresponding to anomalous traffic classes among the full set of detected clusters still remains a challenge. This is mostly due to the nature of clustering techniques that work on traffic samples of a given duration, each cluster being classified after an uncertain post processing stage. In this paper, we show how anomaly detectors can benefit from keeping a temporal track of the clustering results along time. This improvement has been added to ORUNADA (Online Real-time Unsupervised Network Anomaly detection Algorithm) that aimed at providing efficient anomaly detection on high speed networks. This new ORUNADA version-called H-ORUNADA for History-ORUNADA-is then evaluated on a new ground truth, called SynthONTS, that is currently designed to provide a modern and complete dataset with labeled anomaly. H-ORUNADA has also been implemented on Spark Streaming for being able to work on very high speed networks (targeting several hundreds of Gbits/s), and evaluated on the Google Cloud Platform.

140791
Les informations recueillies font l’objet d’un traitement informatique destiné à des statistiques d'utilisation du formulaire de recherche dans la base de données des publications scientifiques. Les destinataires des données sont : le service de documentation du LAAS.Conformément à la loi « informatique et libertés » du 6 janvier 1978 modifiée en 2004, vous bénéficiez d’un droit d’accès et de rectification aux informations qui vous concernent, que vous pouvez exercer en vous adressant à
Pour recevoir une copie des documents, contacter doc@laas.fr en mentionnant le n° de rapport LAAS et votre adresse postale. Signalez tout problème de dysfonctionnement à sysadmin@laas.fr. http://www.laas.fr/pulman/pulman-isens/web/app.php/