Programme Session Posters
Mohammad Abualhoul (INRIA, Paris-Rocquencourt) - Visible light and radio communication for autonomous driving
Fully automated vehicles have the potential to improve the comfort of humans’ life. For driving from one place to another, an automated vehicle must avoid collisions and be able to select non-congested roads for safe and efficient driving. In order to do that, the vehicles need to control the mobility in both macroscopic and microscopic levels by utilising information exchange with other vehicles using dedicated wireless communications technology. Currently, ISO TC204 WG16 has been defining ITS architecture for communications over diverse media including DSRC, IEEE802.11p, and 3G/4G. Even though these medias typically provide high data rate over relatively long coverage range (minimum several 100’s meters), it is expected to experience a channel congestion and low communication reliability, especially for vehicular high dense scenarios. In such scenarios, vehicles still required to control the mobility with maintaining the vehicular safety requirements. Our poster will present the research proposal to look for supportive and at the same time effective communication media with the ability to support sufficient connection between vehicles. The research general overview suggests deploying vehicular lighting systems as Visible Light Communication (VLC) link to exchange information between vehicles, combined with other Radio-based communication solutions for ITS.
Imed Allal (Orange Labs) - A new wake-up mechanism for access equipment in mobile networks based on low power wide area networks
The Radio Access Network (RAN) represents the large amount of the annual energy consumed in mobile networks. The Base Station (BS) constitutes the main component of the RAN. The switch ON/OFF of BSs is one of the mostly used techniques to reduce the energy consumption cited in the literature. With these mechanisms, BSs are in a standby mode (OFF) during low traffic periods. However, the BS continues to consume a quantity of energy during the switch OFF periods, which represents a significant percentage of the amount of energy consumed by the BS during a day. In order to tackle this issue, in this poster, we propose a new wake-up mechanism for base stations in mobile networks based on Low Power Wide Area Networks (LPWAN). Our proposal aims to carry out a complete shutdown of the base stations during low traffic periods in order to reduce the Radio Access Network (RAN) energy consumption. Then, the wake-up process is handled by the Core Network through a LPWAN dedicated for Internet Of Things (IoT) applications such as the Long Rang (LoRa) system.
Lamine Amour (Université Paris Est) - Proposition d'une approche algorithmique de streaming adaptatif basée sur la perception utilisateur (QoE)
Depuis l’avènement de l’internet et l’apparition de la première génération du mobile (1G), il ne passe pas une année sans que les performances soient améliorées afin de répondre aux besoins sans cesse croissants des services multimédias (transfert images, Voix, …etc.). Néanmoins, la forte progression des besoins nous pousse aujourd’hui à nous questionner sur la pertinence de la politique actuelle qui consiste à sans cesse surdimensionner l’infrastructure. L’une des applications les plus consommatrices en ressource à l’heure actuelle est la vidéo à la demande. Nous proposons dans ce poster une approche algorithmique de streaming adaptatif dans un réseau de communication à trafic irrégulier de type IP, basée sur un modèle prédictif de la perception utilisateur (QoE). Cette approche est implémentée dans un lecteur HTTP DASH (Dynamic Adaptive Streaming over HTTP) qui standardisé par le groupe de recherche MPEG (Moving Picture Experts Group).
Mehdi Anteur (IRIT, Toulouse) - Ultra narrow band technique for low power wide area communications
Ultra Narrow Band (UNB) transmission technique makes a powerful comeback in the frame of Machine-To-Machine (M2M) communications thanks to advances in signal processing. Its properties - low power consumption, resistance to interference and long range with small RF power - makes it very attractive for satellite communications. However, because of their small bandwidth, UNB signals are more sensitive to frequency shift and frequency drift. Such effect are usually not taken into account and there is a lack of model to study a system using UNB signals. This is why, after establishing an exact multi-user interference model and showing the limit of classical models for the study UNB signals, we propose a general semi-analytical model to evaluate the performance of a UNB system that takes into account frequency effects. This model is then used to assess the packet loss ratio and the throughput of a representative LEO mobile satellite system.
Sabrine Aroua (L3i, Université de la Rochelle / CRISTAL, ENSI Tunisie) - Allocation non égoïste des canaux dans les réseaux à radio cognitive pour les smart grids
Dans le travail présenté, nous nous intéressons du déploiement des réseaux de capteurs sans fil à radio cognitive pour la fonction smart metering et la fonction demande-réponse des Smart Grids (CG). Au moins un nœud capteur est déployé au niveau de chaque maison pour collecter les données relatives à la consommation et à la production de l'énergie. Une fois les données collectées, elles seront par la suite envoyées vers un nœud central appelé le Sink. Traditionnellement, pour échanger les messages de contrôle, les SU utilisent un canal de contrôle en commun (CCC) partagé entre eux. Ce canal est supposé toujours disponible, libre du signal primaire. Mais l'hypothèse de l'existence du CCC est non pratique dans un environnement réel. D'où, notre objectif est la proposition d'une solution d'allocation équitable des canaux sans l'utilisation du CCC. Notre but est que chaque nœud soit capable de choisir le canal adéquat pour transmettre ses données, tout en respectant les priorités de ses nœuds voisins dans l'accès aux canaux. Le problème de l'allocation des canaux est formulé en une chaine de Markov partiellement observée. Notre solution a été simulée sur OMNeT++ pour mesurer ses performances. En comparant notre approche avec la solution qui se base sur l'utilisation du CCC, la simulation montre que la solution proposée améliore l'utilisation du spectre malgré l'absence du CCC. Ces résultats montrent le non égoïsme des nœuds capteurs et que la solution s'affranchit bien du CCC.
Mohamed Attia (DRIVE, Nevers) - A new intrusion detection approach against lethal attacks in the smart grid: temporal and spatial based detections
The smart grid is the new vision of the traditional power grid, which is characterized by the integration of communication network between all its components beginning from the producers, going through the transmission and distribution units and finishing by the consumers and end users. The heterogeneity of its components imposes a sophisticated security architecture to protect the smart grid from any attack's attempt. In this context, we propose an architecture in which Intrusion Detection System (IDS) agents are implemented with a distributed manner to monitor the consumer side. The IDSs placed in the smart meters help to monitor building appliances and those placed in collectors are used to supervise smart meters. Other IDSs are implemented in a centralized manner at the control center to monitor the aggregators. Those IDSs rely on rule-based detection policy, which consists on the combination of temporal and spatial detection rules. Simulation results prove that this combination enhances attack’s detection rate and reduces false positive rate.
Henry-Joseph Audéoud (UGA - LIG, Grenoble) - Routage efficace et garanti dans les réseaux de capteurs sans fils
Les protocoles de routage (RIPv2, OSPF, BGP, EIGRP…), largement utilisés dans le réseaux Internet ne sont pas adaptés aux contraintes particulières des réseaux de capteurs, pour lesquels le budget énergétique est limité, le canal de communication facilement saturé, la mobilité parfois élevée, où de nombreux liens sont asymétriques… Des protocoles de routage spécialement conçus en tenant compte de ces contraintes ont donc vu le jour, comme AODV ou RPL. Le protocole de routage que nous proposons, nommé LRP (Lightweight Routing Protocol) se démarque de ceux-ci sur deux points. Le premier est l'absence garantie de boucle de routage, quand bien même la topologie du réseau changerait. Le second est que LRP est conçu pour être le plus silencieux possible. Dans ce but, nous utilisons un arbre de collecte construit à l'aide de l'algorithme de Bellman-Ford distibué, ainsi qu'un ensemble de routes d'hôtes, construites soit proactivement soit réactivement. Un certains nombre de règles locales permettent d'éviter les boucles de routage, alors qu'un mécanisme de réparation à bas coût permet de réparer l'arbre avec peu d'échanges de paquets. Afin de proposer un résultat pratique en parallèle des travaux théoriques, LRP a été testé sur la plateforme FIT IoT-lab.
Fetia Bannour (UPEC, LiSSi Lab, Paris) - The SDN controller placement problem
The Software-Defined Networking (SDN) architecture moves the control logic to dedicated software-based controllers responsible for managing the whole network based on a logically centralized view. Such separation between the control and data planes introduces new concerns regarding the control plane scalability, reliability, consistency. In particular, the distributed SDN control plane architecture raises questions about the required number of SDN controllers and their optimal locations for the purpose of enhancing the network performance and therefore guarenteeing the users Quality Of Experience (QoE). The poster investigates and formulates the SDN controller placement problem as a multi-objective optimization problem using the DCLARA clustering algorithm and the Sinalgo simulation and implementation framework. It proposes a novel approach to distributing the SDN control plane based on the desired perfomance metrics and requirements.
Thiwiza Bellache (UVSQ/VEDECOM) - Contrôle dynamique des communications dans un environnement V2V et V2I
Les futurs véhicules de transport de personnes ou de marchandises seront dotés de moyens de communication avancés permettant l'échange d'information aussi bien de véhicule à véhicule (V2V) qu'entre le véhicule et l'infrastructure routière (V2I) et/ou les grands réseaux de télécommunication mobiles cellulaires. La disponibilité de ces moyens de communications évolués permet une gestion efficace du trafic routier tout en offrant à l'utilisateur la sécurité du véhicule et de ses occupants en temps réel tout le long du trajet. Deux classes de services de communication devraient être déployées dans l'architecture de l'ITS. La première est destinée au transport des flux temps réel à contraintes temporelles strictes, la seconde au transport des flux à contraintes temporelles élastiques. De plus l'exigence des flux en termes de débit offert doit être prise en compte. La couche radio préconisée dans une première étape du déploiement de l'ITS repose sur la technologie 802.11p. Dans les évolutions à venir qui accompagnent la migration des réseaux mobiles au-delà de la 3G l'intégration d'une technologie radio dérivée du LTE devrait permettre d'offrir des services de communication large bande aux applications de l'ITS et permettre ainsi l'interopérabilité de ces services avec ceux prévus dans les futurs réseaux 4G/5G. Cette thèse vise principalement l'analyse et l'optimisation des protocoles de transport des flux de données. Les travaux de recherche envisagés se subdivisent en plusieurs axes. L'analyse et la modélisation des applications de l'ITS avec comme point de départ le choix d'un ou plusieurs scénarios de réseau véhiculaire et de classes d'applications qui échangent de l'information entre véhicules (V2V) ou entre véhicule et infrastructure (V2I). Déploiement d'un processus de contrôle distribué au bord des véhicules qui permettra d'une part, de maintenir la connectivité du réseau reliant les véhicules entre elles ou avec l'infrastructure. D'autre part, d'assurer un routage cohérent. Evaluation et minimisation du temps de transit de bout en bout et de taux de perte au sein de l'architecture des protocoles en prenant en compte la mobilité spatiale et l'état de la congestion du canal radio.
Hajer Ben Rekhissa (iQsim / LEAT, Nice Sophia Antipolis) - Réduction de la consommation énergétique dans un module M2M utilisant la technologie LTE-A
De nos jours, la communication des machines est devenue une technologie incontournable, vu les services qu'elle est capable de proposer dans nombreux domaines. Afin d'assurer la communication des modules M2M (Machine to Machine), plusieurs technologies peuvent être utilisées ; des technologies à courte portée tel que le Zigbee, Bluetooth et des technologies de longue portée comme les technologies cellulaires privées (sigfox, Lora, Weightless, Onramp, ...) ou les technologies cellulaires standards (2G, 3G, 4G). De plus, un module M2M, alimenté par une batterie, est généralement déployé dans des zones inaccessibles, d'où vient l'intérêt de choisir une technologie de communication qui assure une longue durée de vie de la batterie et par conséquent une longue durée de vie du module M2M. L'objectif de ce poster consiste à présenter les stratégies de réduction de la consommation énergétique dans un module M2M utilisant la technologie Long Term Evolution-Advanced (LTE-A).
Mariem Ben Yahia (Orange Labs) - A video packets filtering strategy combined with http/2.0
In VoD or live stream delivery, we may deal with situations where the average bit-rate of the video is greater than the available capacity of the network links. This bandwidth shortage results in a drop of video packets at the bottleneck of the link regardless of packet content. However, some video frames are more important than others because they contain more important information. Also, the different video frames depend on each other’s. We suggest a lightweight packet filtering strategy in order to select the set of frames that we can lose without a great impact in the distortion of the quality video when the bandwidth decreases. The strategy consists in attributing an indicator of priority to each video frame regarding its type (I, B, P), its dependencies to other video frames, its size and its display time. A combination of all these criteria determines the evaluation of a frame. HTTP/2.0 offers the possibility to multiplex the requests and the responses by sending sequentially a part of the content that belongs to a special stream. Actually, each stream has an ID stream, the ID of the stream it depends on and a weight indicator defined by the client. In any time, the client is able to change the weight of a stream by sending a PRIORITY frame to the server. The idea is to send every video frame in an HTTP/2.0 stream and to schedule video frames taking into account the bandwidth performances, the importance of the contents and the video frames display times.
Younes Bouchaala (Paris Saclay / VEDECOM / INRIA Paris) - Utilisation de la communication véhiculaire pour l'amélioration de la sécurité routière
L'émergence de la norme IEEE-802.11 et des réseaux locaux sans fil a fourni une technologie de transmission radio assez fiable et relativement peu onéreuse de sorte que la difficulté technique pour les réseaux véhiculaires n'est plus au niveau physique mais sera plutôt au niveau des protocoles. Dans un réseau VANET (Vehicular Ad Hoc Network), le trafic de sécurité se base principalement sur l’échange périodique de messages de signalisations CAM (Cooperative Awareness Message), et occasionnellement des messages événementiels DENM (Decentralized Environmental Notification Message). Etant assez fréquents, les messages périodiques de sécurité, i.e. CAM, ont tendance à saturer le canal radio, ce qui revient à réduire l’ensemble des performances du réseau. Nos études consistent à optimiser l’exploitation de la bande passante dans les VANETs, plus précisément à optimiser les protocoles et les mécanismes du standard ITS-G5 (le standard Européen de la technologie IEEE 802.11p) considéré comme le média principal des communications véhiculaires. Actuellement nous étudions les performances des transmissions dite One-hop dans un réseau VANET et nous tâchons d’optimiser le débit échangé entre un nœud et son plus proche voisin vu que ce dernier est le premier concerné par les messages broadcastés par le nœud émetteur. Nos résultats obtenus, théoriquement et par simulations, montrent qu’il est possible d’optimiser l’exploitation de la bande passante, i.e. optimiser le débit échangé avec le voisin le plus proche, en optimisant les paramètres de communication en fonction de la densité du réseau. Les prochaines étapes de nos travaux consistent à développer un algorithme de décision dynamique qui fera varier les paramètres de communication en fonction de la variation des caractéristiques du réseau (e.g. la variation de la densité des nœuds).
Pierre Brunisholz (UGA - LIG, Grenoble) - City-Wide Wi-Fi coverage
Smartphones have been massively adopted by the public. These devices get more and more sensors (such as GPS, temperature sensors and accelerometers), and are connected through standard technologies (WiFi, BLE, LTE, ...). This make them perfect devices for ubiquitous data collection, as well as relay for sensors with no direct Internet access. As there is a lot of WiFi Access Points in today's cities, we want to evaluate the possibilities offered by the current coverage in terms of data transportation. A first issue is related to the Access Points location. Even if the number of public Access Points is growing, they are deployed in specific locations and a lot of areas remain uncovered. If we consider that private Access Points could dedicate a certain amount of bandwidth for data collection purposes, we can obtain a near city-wide coverage. However some places such as bridges and parks will remain uncovered. A second issue, is that we are considering moving users and that WiFi is not built to handle mobility. The whole handover process can take a long time and is not fit to work with high speed users. These two issues introduce gaps in connectivity. In order to have an idea of these connectivity patterns, we simulate mobile users moving at different speeds and with different handover values. This simulation was carried out in Grenoble with Access Points coordinates corresponding to Freeboxes and topologically generated coordinates.
Hugo Chelle (IRIT / Airbus Defence And Space, Toulouse) - Modelling discontinuous LEO satellite constellations for M2M applications
Machine-To-Machine (M2M) is the rising up market in telecommunications. The application scope of M2M is widespread, there are applications from the local environment as RFID-tagged livestock on a farm to worldwide applications as the airplane tracking. Discontinuous LEO satellite constellation are expected to be more and more exploited in future satellite telecommunications systems especially in the context of delay tolerant network such as messaging and M2M communications (e.g. freight/plane tracking, wind turbine sending their daily energy production report). M2M communications shall be the lowest energy consuming. As the satellite coverage is discontinuous, the satellite access is a critical part for system sizing. Therefore, the modelling of such kind of constellation and the traffic modelling become essential in order to realize protocols performance comparison. At the best of our knowledge, no modelling to study protocol performance for this kind of constellation has been proposed yet. We highlight two different points of view of study: satellite and ground. In this poster, we widely analyse the satellite point of view and define a new traffic modelling for M2M communications at the satellite point of view. Because of the satellite discontinuity, the classical traffic modelling for terrestrial or satellite networks are inadequate. The poster explains the generation of new messages to transmit under the satellite coverage. To illustrate, the proposed model is used to generate new message and to compare the performance of two different MAC protocols using the same physical layer.
Guangshuo Chen (INRIA, Saclay) - On the quest for representative behavioral datasets: mobility and content demand
Mobile datasets are widely used as firsthand sources for human mobility research. These datasets are often incomplete or have heterogeneous spatiotemporal resolutions, e.g. a dataset is often aggregated or in lack of fields. In many cases, a reliable dataset in human mobility research comes from sampling or merging original datasets, a challenging task. In this paper, we present our experience on creating a reliable dataset describing mobile data traffic in individual’s spatiotemporal view. We focus on individuals having enough geographical information and merge their call records from one dataset with the data traffic records extracted from another dataset. Based on this dataset, we perform an analysis of user demand on mobile data traffic in terms of spatial and temporal behaviors. For each subscriber, sessions are put into a 3-dimensional space in terms of space, time and volume and are clustered by applying DBScan. Characteristics of are revealed from the statistical analysis on clusters. Subscribers are also categorized according to their clusters.
Timothy Claeys (IMAG, Grenoble) - Energy-efficient rejoining scheme for TSCH networks
The old IEEE 802.15.4 standard does not meet the requirements to support the emerging IoT. The IEEE 802.15.4e standard adds two new MAC-layer protocols. The first one is the Deterministic and Synchronous Multichannel Extension (DSME) which uses the same beacon format as the old standard beacon enabled PANs, but adds channel hopping, which makes it more robust to interferences. The second new protocol is TSCH. TSCH describes a medium access technique which uses time synchronisation to achieve low-power operation and channel hopping to enable high reliability. An initial time synchronisation of the network is performed during the bootstrap phase. All the clock systems of the participating nodes are calibrated and the nodes receive a slotted schedule which imposes time and frequency synchronization. TSCH can be used in energy scavenging networks where the main issue is the scarce energy supply. Networks that use energy scavenging have to take into account that the energy harvested from the environment can vary very strongly as a result of environmental factors. Due to the uncertainty in energy supply, nodes may frequently drop out of the network and return later when their batteries are sufficiently charged. The current TSCH standard does not propose a solution for resynchronization of lost nodes in existing networks. Discovery of lost nodes should proceed as quickly as possible to allow the network to resume its optimal operation. Fast discovery helps avoiding long listening periods, and consequently, helps improving energy efficiency. Our research proposes a new resynchronization scheme that minimises the energy consumption and join latency. The algorithm is implemented in the OpenWSN stack and tested on the TelosB, Greennet and OpenMote platform.
Yoann Couble (IRIT, Toulouse) - Resource management optimization in the return link of satellite communications
Satellite networks are becoming more and more relevant to provide internet services to areas where deploying cellular or cabled networks is not commercially viable, areas which represent approximately 3 billion people accross the world. Initially only launched for TV broadcasting services, field where they are still thriving, the developpement of more performant antennas (especially multi-beam antennas) and the use of the Very High Frequency Band (Ku at 20GHz and Ka at 30GHz) has now made GEO satellite networks a serious competitor to other technologies. The use of multi-beam antennas introduces co-channel interferences to be accounted in the link budget. In this poster we present the main principles of interference management in the forward and the return link of a transparent satellite, and present a few classical static schemes used in cellular networks as well as satellite networks. Then, we present the main challenges of our work regarding dynamic resource allocation in the return link.
Khaoula Dhifallah (Orange Labs) - Gestion de ressources réseau partagées dans un contexte de crowd networking
Pour répondre au challenge de la montée en débit (d'un ordre x1000 entre 2010 et 2020), de nouvelles pistes sont considérées tels que le déploiement de petites cellules (small cells). Mais ce postulat d'ordre technologique est confronté à une contrainte terrain forte d'ordre économique, où les opérateurs ne peuvent pas assumer à eux seuls les coûts de site liés au déploiement de ces nombreuses petites cellules. Ces coûts seront partagés via de nouveaux modèles d’affaire et de ruptures technologiques s'y adaptant. Ainsi, le déploiement de petites cellules pourra contribuer à la fois à la montée en charge progressive de l'Internet dans les pays émergents et à la densification dans les pays industrialisés. La miniaturisation des équipements et l'accroissement des capacités des terminaux contribuent à l'essor du modèle "crowd networks" où l'opérateur déploie des petites cellules sur des infrastructures existantes qui appartiennent à des partenaires industriels ou des collectivités locales. La réduction de coût provient de la mutualisation entre plusieurs domaines. L'enjeu dans les pays industrialisés est l'évolution vers la très forte densité, dans un contexte multi-acteurs. La première étape de la thèse est d'étudier le contexte du sujet et de faire une modélisation du réseau afin de comparer les différentes politiques de partage de ressources et de responsabilités entre opérateurs et "prosumers".
Saad El Jaouhari (Télécom Bretagne, Rennes) - Innovative usage of WebRTC for the Web of Things
Due to the IETF and W3C efforts, a new real time communications technology called WebRTC appeared. WebRTC provides the ability to exchange media directly between peers just with theirs browsers, which means no need to download plugins or a software to establish a communication session. Currently, many browsers such as Google Chrome or Mozilla Firefox natively support WebRTC. Beside the powerful communication capabil- ities provided by WebRTC, security was considered from the beginning in order to guarantee authentication, confidentiality and integrity in the system. Moreover, WebRTC allows the establishment of secure and reliable data channels between the communicating parties, which open the consideration of very rich scenarios, such as exchanging data, during an ongoing We- bRTC multimedia session, involving smart objects at the end sides of the session. Those smart objects ( sensors and / or actuators) can be located either on the caller’s side or on the callee’s side. Such scenario can be applied to several fields such as in Smart Home, in ITS (Intelligent Transportation Systems) and in Smart E-Health. This approach allows us to interconnect two worlds (WebRTC and WoT) with the Web as a native support. Web of Things (WoT) is by the way the application of the existing and the emerging tools and techniques used on the Web in the development of different innovative IoT scenarios. The advantage of using such approach is that we can reuse the available and widely popular Web protocols without the need of reinventing another complex protocol. Moreover, developers can write applications that can interact with the smart objects in the same way as with any other Web service, in particular using the RESTful architectures. Concretely, Things can have a public URL, accessible through the Web. We are interested in this thesis in the development of innovative uses of WebRTC as part of the Web of Things. Such uses introduces several issues, mainly on: 1) The security and privacy related problems. Obviously, only authenticated and authorized entities are allowed to interact with the smart objects. Moreover, this interaction needs to be secured using an appropriate protocol. Trust and privacy problems are also on of the main interest of this thesis. 2) The synchronization problems, since the media and the data are transmitted using different secure channels, a certain level of synchronization need to be provided depending on the context and the use cases. For instance, in case of an accident, where a paramedic is communicating remotely with a doctor and sends him data coming from a heartbeat sensor attached to the victim, strong synchronization between the media and the data is required.
Jean-Gabriel Krieg (IRIT, Toulouse) - SmartPark: unlocking the smartphone's senses for smart city parking
Studies have shown that in places like New York City drivers often spend over 20 min looking for parking, contributing to as much as 30% of the total traffic. In response, cities like San Francisco have deployed systems capable of pointing drivers to the closest available parking spot. Unfortunately, such systems have gained little traction as they rely on specialized infrastructure that is expensive to build and maintain. We present SmartPark, a smartphone based system that relaxes the requirement for specialized infrastructure by relying on the smartphone’s sensors and ubiquitous Wi-Fi and cellular infrastructure. To accomplish this, SmartPark addresses two major challenges, under the constraint of minimum impact on battery life: automatic transportation mode detection and location matching. Solved together, they enable SmartPark to automatically detect when a user pulls out of a parking spot, making it available again.
François Lemercier (Telecom Bretagne, IRISA) - Support for hybrid network in the RPL routing protocol
Advanced Metering Infrastructure AMI is a smart grid component that provides two-way communication between providers and consumers, for example, to inform customers on electricity prices or perform utility management. AMI mostly relies on a multitude of devices using multiple heterogeneous technologies, such as wireless or power line communication. Especially for redundancy purposes, this heterogeneity has to be taken into account by the routing protocol. The IETF ROLL Working group defined a distance vector routing protocol, called RPL, that organizes a topology using Directed Acyclic Graphs (DAG). RPL becomes very popular in the Internet of Things area, and is considered as the based protocol in smart grids. However, its support for hybrid network is limited. Our work consists in designing innovative approaches to take advantage of multiple interface devices in RPL. To do so, we investigate three alternatives to manage the interface selection policy. The first solution consists of using the multiple instance characteristic of RPL, using one instance per technology. For the second solution we combine the link quality of all the interfaces and let RPL choose the best neighbor according to this new metric. The third solution sees an interface as a neighbor and thus a node will select the best couple (node, interface) as a parent in the DAG. Using an ad hoc Python simulator, we present some preliminary results, which highlight the benefit of using multiple interfaces.
Andreina Liendo (IMAG, Grenoble) - Efficient Bluetooth low-energy operation for low-duty cycle applications
Bluetooth Low Energy has been designed as a power efficient protocol for small portable and autonomous devices like wearables and smart watches. BLE shows its efficiency for connecting these devices with smartphones to exchange data periodically and frequently, like heart rate or notifications. Nevertheless there is a need for adapting the protocol in order to extend its applicability to more generic IoT scenarios, where traffic patterns are different and lifetime should reach many years. BLE has been shown to be more energy efficient and robust than other well known WSN protocols, which makes it strong candidate for its use in IoT related scenarios. In addition, BLE is present in almost every smartphone, which turns them into perfect ubiquitous remote controls for smart homes, buildings or cities. In this paper we propose new operating modes for BLE to provide a much wider applicability range covering asynchronous, aperiodic and very low frequency scenarios while extending overall coin cell battery lifetime, without modification of the existing BLE specification. We show how to efficiently adapt the protocol for different IoT scenarios depending on their requirements, thus optimizing the use of the battery capacity as much as possible while guaranteeing acceptable latency. Our results on two typical test-cases show that by using an optimized configuration and appropriate operating mode based on scenario requirements, lifetime can be increased up to 1.8× (4.1 to 7.3 years) for a BLE master in a first case, and 58.3× (3.05 months to 14.82 years) for a BLE slave in a second case. Both cases represent typical scenarios, showing that BLE can be made compatible with a vast amount of IoT-related use cases.
Diogo Mattos (LIP6, Paris) - A lightweight protocol for consistent policy update on distributed software-defined networking
Network policies define the proper behavior of software-defined networking as forwarding traffic throw network functions and middle boxes. Policy updates, however, arise from all network controllers and they have to be committed in a consistent way. Otherwise, the network may experience unexpected transitory configuration states, which compromise its performance, security or, even, operation. In this work, we propose a consistence protocol to serialize policy updates and to compose network policies avoiding conflicts. The main contributions are: (i) a consistence protocol to serialize commitment of policy updates; (ii)~a consensus interface to controllers agree with the version of the network configuration; and (iii) an algorithm for checking if new policy is an update, a refinement, or if it conflicts with other policies already installed. We prove that our consistence protocol achieves a global order for all policies updates, and our algorithm correctly composes all policies. Our analyses show that the proposal achieves a per-packet consistent policy update at a low overhead cost.
Salma Matoussi (LIP6, Paris) - SDN based virtual RAN
The project deals with future 5G mobile networks, using the Software Defined Networks (SDN) and Virtual Radio Access Network (VRAN) concepts. SDN is a new networking paradigm facilitating network programmability and network management. It decouples the control plane from the data plane in network equipment, transforming switches and routers into simple forwarding devices that apply rules sent by a remote controller using a standard protocol. On the other hand, VRAN offers the capability of creating multiple instances of RANs upon the same physical equipment and calibrating their capacities on demand. Each instance of VRAN will be associated to one Mobile Virtual Operator (MVO), this way offering the RAN as a Service (RANaaS). Accordingly, available physical resources are exploited by different MVOs, which enhance the revenue of the VRAN provider. Thanks to the virtualization technology, the isolation between VRANs is guaranteed and hence interference probabilities are minimized. In this context, the thesis research focuses on orchestrating virtual resources (i.e., virtual radio resources, virtual base band units (vBBU), processing power, CPU, memory, network interfaces, etc.) leveraging the SDN concept. From the algorithmic point of view, the challenge will be to design, model, and analyze efficient resource allocation algorithms for the VRAN provider, taking into account several parameters such as, its revenue, the Quality of Service of end-users, energy consumption, and resource availability. Consolidation of virtual machines (i.e., strategic deployment of vBBU for an efficient usage of network resources) is also a challenge to tackle. Indeed, designing and implementing an optimal algorithm in the SDN controller that first consolidates virtual machines (vBBU) to minimize the number of active base stations and, in a second stage, the set of active links and virtual remote radio heads (vRRH) to be turned ON is an interesting avenue to further enhance the performance of virtual RANs.
Abdoul Aziz Mbacke (INRIA, Lille) - Distributed, efficient and fair anti-collision solutions for dense and mobile RFID deployments
Radio Frequency Identification or RFID is a technology that allows for con- tactless and without line of sight identification of goods and/or individuals. While this task has been accomplished thanks to barcodes and laser readers for decades, the advent of electromagnetic technologies allowed for improvements in this regard. As such, RFID technology grants a device name reader the pos- sibility, thanks to a technique known as "backscattering", to interrogate other smaller devices named tags. Given the dimensions and characteristics of tags, they can be attached to goods or individuals, in order to retrieve previously stored information. RFID thus serves as an impulse to create new applications where tags and readers are deployed in large numbers in order to track the position of goods, improve security, enhance traceability, smarter cities, etc. However, such large deployments come at an expense inherent to any wireless communication: collisions. Indeed, having a large number of promiscuous read- ers operating at the same time induces a lot of interferences that, in the case of RFID, translates into unread tags, sluggish operations and inefficient energy consumption. Our poster presents DEFAR (Distributed, Efficient Fair Anti-collision for RFID), an algorithm that allows consistent improvements in regards of collisions reduction, efficiency and fairness compared to the best algorithm in the literature. We also introduce two proposals intended for dense and mobile deployments of RFID readers with steady results in terms of coverage delay.
Dianne Scherly Varela de Medeiros (LIP6, Paris) - Spread betweenness centrality: why quasi-shortest paths matter
Betweenness centrality metrics based on shortest paths completely neglect the contribution of nodes that are still central, but mostly participate in quasi-shortest paths, which are slightly longer than the shortest ones. The idea of this work is to consider also such not so central nodes, assigning them more importance. To this end, this work proposes the spread betweenness centrality (ρ-betweenness), which considers both shortest and quasi-shortest paths, assigning weights to each contribution based on how longer is the path on which the intermediary node falls compared to the shortest one. This weight depends on the spreadness parameter ρ, that determines the maximum tolerable difference between the lengths of the shortest path and the quasi-shortest path. The metric analysis is formalized through the comparison between the results of the spread, traditional and distance scaled betweenness centralities, using four datasets. Results show that the ρ-betweenness captures nodes that were once ignored by the other centrality metrics, being able to identify poorly classified nodes and, also, to assign more importance to central nodes not necessarily in shortest paths. The proposed metric is capable of breaking ties between nodes that in fact are not equal and it can avoid reallocation of network resources.
Mohamed Ayoub Messous (Drive Lab, Université de Bourgogne) - Efficient mobility model for a fleet of interconnected UAVs
The main focus of the work that we are presenting is on developing an original distributed mobility model for autonomous fleet of interconnected UAVs (Unmanned Aerial Vehicles) performing an area exploration mission. The UAVs, equipped with wireless ad-hoc capabilities, are required to optimally explore an area while maintaining connectivity with their neighboring UAVs and the base station. These two metrics are reasonably important in a wide span of applications, wherever achieving the best area coverage and maintaining network connectivity, amongst UAVs, is a key feature. Nevertheless, another equally noteworthy constraint should be joint to these two metrics, Energy. Because energy is a scarce resource, especially for UAVs, its wise management is quite beneficial for the network lifetime and mission success. Hence, the proposed mobility model, compared to other models in the literature, is the first to ever include the remaining energy level as decision criterion combined with area coverage and network connectivity. Based on these criterions and the information received from its neighbors, each UAV determines its next movement to be undertaken. The performances of the proposed approach are compared with those achieved through a randomized approach and a forces-based approach. Simulation results, using NS3, show that it outperforms the two other models in terms of coverage and connectivity.
Yuqi Mo (CITI Labs, Lyon) - Analysis and optimization of the capacity for UNB transmission based IoT networks
IoT (Internet of things) networks are gaining more and more interest as they will provide connectivity for countless number of devices. When the goal is to collect information (for example, for temperature monitoring, parking availability monitoring, smart metering, ...), the devices will require to transmit only a few amount of data, during a small proportion of time. In this case, the base station should cover as much devices as possible to reduce the cost of infrastructure deployment, and not waste transmission resources. This is the reason why long range low power transmission technology as UNB (Ultra Narrow Band, known as SigFox) has emerged recently. In a UNB based network, devices access to the medium at will. Transmissions are realized randomly, both in time and frequency (R-FTMA : Random Frequency and Time Multiple Access). This channel access scheme saves the resource for reservation, but it also induces some simultaneous transmission in the same frequency area, leading to interference. The system performance of such network was evaluated by previous work. To ensure the reliability of the UNB based network, the poster proposes a mechanism of replication, as well as a study of the impacts of the replication numbers. Theoretical analysis and results of validations for the network's performance has been provided. Results are exploited to find out the optimum number of replications as a function of a unified network configuration and to estimate the minimum number of replications for a targeted network performance.
Elodie Morin (UGA - LIG, Grenoble) - Interoperability in WSN - Understanding WSN technologies features
Ultra low energy consumption is critical for the Internet of Things, since most of the devices need to operate on batteries and offer long lifetimes on the order of several years. For energy harvesting nodes that get the energy from the environment, the energy consumed during their operation needs to be balanced with the energy intake, which usually remains very small. The estimation of energy consumption is thus crucial to choose the right technology and configuration parameters among several available standards for wireless IoT communications. We will present a tutorial on the energy consumption of the main wireless standards: IEEE 802.15.4/e, Bluetooth Low Energy (BLE), IEEE 802.11 Power Saving Mode (PSM), IEEE 802.11ah, and new emerging long-range technologies such as LoRa and SIGFOX. Our analysis computes the energy consumption based on the time required to stay in a given state (Sleep, Idle, Tx, Rx) for each protocol, by considering PHY and MAC layers, and assuming IPv6/6LoWPAN traffic whenever possible, perfect transmission conditions, along with no contention. We carefully analyze the energy consumption patterns for duty-cycled MAC (Medium Access Control) layers to derive the lifetimes of a battery-operated node that generates data and uploads them to a main-powered sink (uplink traffic). The poster presents the results of this analysis.
Raphael Naves (Thales Communications & Security) - Physical layer network coding in Ad-hoc networks
We investigate the potential gains of Physical Layer Network Coding (PLNC) in large multihop topologies. Physical Layer Network Coding concept has been first introduced as a solution to increase the throughput of a two way relay channel communication. Unlike most wireless communications techniques which try to avoid collisions, the main idea of PLNC is to send simultaneously two messages to a common receiver which are summed at signal level then decoded at packet level. In basic topologies, Physical Layer Network Coding significantly enhances the throughput performance compared to classical communications. However, the contribution of PLNC in large Ad-hoc networks is still an open question. We therefore study the impact of this paradigm on the achievable throughput in realistic radio deployments. Our theoretical approach exploits Linear Programming to derive optimal routing and scheduling schemes in these studied topologies. Our work highlights the important gains that can be achieved when packets are optimally scheduled for transmission in any considered network. Such results set theoretical benchmarks for designing new access methods and routing protocols to efficiently exploit the Physical Layer Network Coding concept.
Kaouthar Ouali (LIP6, Paris) - Towards modeling D2D handover in 5G cellular networks
Device-to-Device (D2D) communication is one of the promising and revolutionary technologies in 5G cellular networks that enables mobile terminals to communicate directly without the base station involvement. One of major objectives of the 5G for this paradigm is to provide seamless mobility satisfying QoS requirements. Hence, the handover process between various cellular networks or more specifically between radio cells is an important issue faced by D2D candidates. In this context, seamless connectivity needs to be maintained when the control plane for one or both of the devices goes through a handover. During this latter process, the two devices have to deal with some problems such as cell selection, the choice of the best time to initiate the handover, a seamless handover, etc. in order to satisfy the D2D communication requirements. In this poster, we mainly focus on the prediction of wrong decision and handover failure probabilities during a handover process in 5G cellular networks in a D2D communication. We propose an analytical model for successful handover probability, missing and unnecessary handover probabilities based on the bandwidth, Received Signal Strength (RSS) and the mobility pattern of the D2D candidates that move with the RPGM model as performance criteria. A queuing system and a Markov chain model are introduced to capture the D2D pair movements between different radio cells.
Jad Oueis (CITI Labs, Lyon) - Autonomous organization of PMR/LTE networks
Deployable PMR/LTE networks are usually required in locations where no operating network exists (e.g. following a natural disaster that damages the network infrastructure), or in out of coverage locations where a network infrastructure does not exist at all (e.g. forest fire or underground rescue). In both cases, the installation of base stations is required. However, with the absence of infrastructure, the base stations do not have a backhaul connection to an Evolved Packet Core (EPC), and are considered isolated. In this case, base stations connect instead to a local EPC. The local EPC must be able to provide the needed functionality for the deployable base stations. Thus, it must include at least the MME, SGW, PGW and HSS functionalities. All the deployable base stations can be co-located with their own local EPC. However, the base stations in the network are served by one local EPC, selected by the base stations as their preferred local EPC. Our contribution is to determine the location of the local EPC to which all the other base stations must connect. In determining the EPC placement, we take into consideration the links between the base stations, their capacity, and their quality, in order to ensure that a given amount of signaling and data traffic can circulate in the network from the base stations to the selected local EPC, and vice-versa.
Philippe Pittoli (ICube, Strasbourg) - Optimisation de DTLS pour les environnements contraints
Transport Layer Security est un protocole défini par l'IETF pour sécuriser les communications sur Internet, et Datagram TLS est sa version utilisée dans l'Internet des Objets. Cependant, le protocole DTLS n'a pas été conçu pour des appareils contraints en mémoire, en taille de code et en vitesse de calcul comme nous pouvons retrouver dans ce domaine. Le protocole ne prend pas en compte en particulier le coût de la connexion en nombre de messages ni la taille limitée des messages échangés dans ces réseaux. Nos travaux proposent une version optimisée de DTLS avec pour objectif la réduction du délai de connexion et du surcoût par message de DTLS durant l'échange de données. Ces travaux fournissent un protocole de communication sécurisé basé sur une connexion plus efficace pour le domaine des environnements contraints.
Ronnie Muthada Pottayya (FEMTO-ST, Besançon) - Adaptability of multimedia streams applied to medical telediagnosis : mobility and data exchange
Over the past decade, due to rapid evolution of device capabilities and communication networks, distributed applications embedding videoconferencing solutions have multiplied. Although much more constrained, the field of telemedicine has not escaped this phenomenon. In the medical area, safety is mandatory. So, to be used in different medical structures (hospitals, clinics, healthcare centers ...) distributed applications, such as telediagnosis for example, must be able to cross the security barriers (secured gateways like proxies, firewalls ...). In this context, User Datagram Protocol (UDP), which is classically recommended for videoconferencing applications (due to real time data), does not cross firewalls unless explicitly configured fixed ports are declared. Indeed, the security policies of these medical centers are not compatible with fixed ports allocations (considered as a security breach). We propose a new video adaptation framework, crossing security gateways, based on transcoding which works, in a very efficient and original way; on TCP (Transmission Control Protocol). This framework is composed of Adaptation Proxies (APs), which have been designed to take into consideration health professionals preferences, network architectures, and mobility of users. The framework can automatically adjust adaptation and communication strategies and uses TCP for the delivery of videoconferencing data. The main advantage of using this protocol is that it crosses easily firewalls and proxies (use of port 80 or 443). However, as TCP is a reliable transport protocol, we needed to design and to employ new intelligent adaptation strategies together with data transmission and reception in order to cope with latency issues and sockets timeout.
Anjalalaina Jean Cristanel Razafimandimby (INRIA, Lille) - Performance assessment in Internet of robotic things
Internet of Robotic Things (IoRT) is a new concept introduced for the first time by ABI Research. Unlike the Internet of Things (IoT), IoRT provides an active sensorization and is considered as the new evolution of IoT. This new concept will bring new opportunities and challenges, while providing new business ideas for IoT and robotics’ entrepreneurs. In various IoRT applications (e.g: smart agriculture, smart environment monitoring, smart exploration, smart disaster rescue, etc), the use of mobile robots’ teams brings many advantages over one powerful IoRT robot. To carry out cooperative tasks, IoRT team members need to communicate with each other, often via a wireless link (i.e. Wifi, Bluetooth). Maintaining communication among multiple mobile IoRT robots is therefore a crucial issue. Our main goal in this work is to control the mouvement of IoRT robots while maintaining the global connectivity to a desired quality-of-service (QoS) level. We have proposed two solutions till now : (i) centralized and (ii) distributed approaches. The centralized approach is based on the computation of the algebraic connectivity and the use of virtual force algorithm while the distributed approach is based on a trained neural network.
Dorin Rautu (IRIT, Toulouse) - VoIP calls and multimedia services with UAV ad-hoc network
Starting somewhere around 2013, a new trend emerged in the toy tech and aerial imaging market, an explosion in popularity of compact multi-rotor aircrafts. We call them drones, UAVs. Since then, the prices lowered and the interest raised, especially in the scientific community. Now, as they are capable to embed quite a payload, we thought, what if we build a network capable of streaming aerial videos, reading sensor data and providing VoIP calls. All with drones, no infrastructure. By deploying this network, we manage to provide VoIP communication services to rescue teams in case of cell coverage blackout or no coverage in the intervention site. But that’s not all. Teams will be able to monitor the progress by the live video feed coming from the drones as well as having additional information collected by the UAVs from sensors.
Alassane Samba (Orange Labs) - Throughput prediction in cellular networks
In cellular networks, Quality of Service, throughput in particular, depends on user context (radio channel quality, speed, distance from base station, etc.). To enhance the Quality of Experience, content providers implement adaptive delivery strategies, where the quality and the characteristics of the delivered content are adjusted to match the QoS of each user. Yet, the content providers take some key decisions at the beginning of the delivery. For instance, most web services have several style sheets for their web page, with a variable number of elements and information. The decision of which style sheet to deliver should be taken in the early moment of the connection although no past throughput observation is available. This calls for new mechanisms to provide a rough prediction of the throughput for the next x seconds, using only contextual information. Existing solutions present several shortcomings preventing their widespread usage. For instance, a well-studied approach is to estimate the bandwidth by a series of short path measurements, including Round Trip Time, and packet loss rate, but this approach requires exchanging data before making a decision. Some proposals rely on instant Channel Quality Indicator to estimate the instantaneous bandwidth, but they do not target throughput prediction on a larger time frame. We aim to identify which contextual parameters are the most relevant to predict the throughput during a session in a cellular network. We have conducted a large-scale trial where users have performed a file download test a thousand times in several locations, mobility and radio connection configurations. Data related to context of use, radio access network performance and traffic quality have been collected for each test. In this poster, we describe this trial and we provide some results of the statistical correlations between the main contextual information and the actual throughput.
Farah Slim (Télécom Bretagne, Brest) - Virtual machine placement in virtualized infrastructure
The virtualization technology is a key enabler of cloud computing where applications are not hosted on dedicated servers anymore, but instead in a number of running virtual machines (VMs) and sharing physical resources of the physical machines (PMs) behind. Thus, cloud-providers must implement efficient strategies to allocate VMs in a flexible way in order to permit the maximum usage of physical resources. To achieve this goal, we need a complete solution of deployment and management of virtualized infrastructure like openstack. The core process of this solution is the resource scheduler; its main role is to allocate resources based on users’ needs leading to overall better utilization of servers’ resources in the datacenters. Openstack is one of these cloud management solutions. This poster investigates first some existing strategies for VM placement, and then the strategy adopted by Openstack pointing out anomalies that could be uncovered within these strategies.
Veronica Quintuna (Orange Labs) - Performance analysis of processing radio access network functions on general purpose computers
The performance analysis of 5G networks is based on the processing time of software-based network functions which are executed on centralized and virtualized eNodeBs. There are several factors that impact the processing time of software-based network functions and/or Virtual Network Functions (VNF); among the most relevant are the compute platform, the operating system, and the virtualization technology. In this way, a predictable performance VNFs becomes a major challenge. The processing time of VNFs and more concretely the performance of scheduling algorithms needs to be evaluated due to real-time constraints present in the base band processing of the Radio Access Network (RAN). This work is focused on the performance analysis of RAN functions executed on General Purpose Computing (GPC) structures and more concretely on the performance of scheduling algorithms for multi-processor systems. We define a scheduling algorithm as optimal if it is able to execute all tasksets of the Base Band Unit (BBU) respecting their deadlines. The computing infrastructure can be sizing according to the CPU utilization by UE and the traffic load handled by the group of centralized eNodeBs. The random nature of processing load by UE and the traffic variability by eNodeB enable both resource pooling and statistic multiplexing in order to save energy and computing resources.
Marion Vasseur (Télécom ParisTech) - Détection de comportements anormaux dans les systèmes de transport intelligents
Les communications dans les systèmes de transport intelligents (ITS) coopératifs peuvent être sujettes à de nombreuses attaques. Le but d’un attaquant peut être de provoquer un accident, de modifier la densité du trafic à son avantage, ou encore d’accéder à des données personnelles comme la localisation d’un véhicule. La sécurisation des échanges dans les ITS coopératifs se fait au travers d’une PKI (Public Key Infrastructure), qui permet l’authentification des véhicules au moyen de certificats. Cependant, il est possible qu’un attaquant possède des certificats valides et soit donc capable d’exploiter les communications pour lancer son attaque. Il est donc nécessaire d’intégrer des méthodes de détection de comportements anormaux au sein de l’architecture C-ITS. Mon travail de thèse consiste à concevoir ces méthodes et de travailler sur une autorité centrale appelée « Misbehavior authority » (MA). Lorsque des stations ITS détectent un comportement anormal, elles envoient des rapports d’accusation à la MA. La MA traite ces rapports afin de décider de l’occurrence du comportement anormal. La station ITS jugée coupable par la MA voit alors son certificat révoqué, l’empêchant de communiquer à nouveau dans le réseau.
Shuai Yu (LIP6, Paris) - Coalition-based energy efficient offloading strategy for immersive collaborative applications in femto-cloud
Computation offloading manages resource-intensive and interactive applications on mobile devices where much processing is replicated with multiple users in the same environment. In this paper, we consider the scenario that multiple mobile users offload duplicated computation tasks to a set of nearby Femto-Cloud called Small Cell cloud enhanced e-NodeB (SCceNB), and share the computation results among them. Our goal is to find an optimal offloading and sharing strategy to minimize the overall energy consumption at the mobile terminal side. To this end, we propose a cooperative call graph to model the problem. Based on the derived call graph, we present a distributed algorithm that combines notions from 0-1 programming and coalitional game to solve it, while considering the delay constraint for each mobile user as well as the computation ability and memory constraints of each SCceNB. Simulation results show that our proposal can reduce energy consumption up to 39.73%, 34.37%, and 19.54% compared to the total offloading scheme, the "no offloading" scheme, and the "optimal offloading without sharing" scheme, respectively.