Avec l’arrivée massive des technologies sans fil, le nombre de terminaux mobiles n’a cessé de croître, pour des usages et des ressources de communication diversifiés. En intégrant les objets du quotidien, nos réseaux de communications sont devenus dynamiques aussi bien en termes de ressources que de topologie physique, offrant accès à des informations de plus en plus riches. La tâche de gestion s’est ainsi complexifiée et requiert des temps de réponse de plus en plus courts difficilement réalisables par un administrateur humain. Il devient indispensable de mettre en œuvre des capacités de gestion autonomes pour les nouveaux réseaux. Dans tous les cas, la gestion d’un système implique une étape essentielle : sa mesure et sa supervision. Peu importe sa nature, c’est cette étape de prise d’information qui permet sa caractérisation, son analyse et son contrôle. Le domaine des réseaux n’échappe pas à cette règle et les objets qui le composent auront besoin d’acquérir des informations sur leur environnement pour mieux s’y adapter. Dans cette thèse, nous nous intéressons au partage efficace de ces informations de mesures à des fins d’auto-analyse et d’évaluation distribuée de la performance. Après avoir formalisé le problème de la mesure distribuée, nous nous consacrons dans un premier temps à l’organisation des échanges de mesures dans les graphes dynamiques. Nous proposons une nouvelle heuristique pour le consensus de la moyenne qui converge plus rapidement que celles de l’état de l’art. Dans un second temps, nous considérons des topologies plus stables pouvant utiliser des flux TCP comme moyen d’échange. Nous proposons un mécanisme d’ordonnancement de ces flux qui conserve le même comportement face à la congestion, tout en réduisant leur latence moyenne. Enfin, nous nous intéressons à l’information de mesure échangée. Nous montrons comment les nœuds peuvent superviser diverses métriques telles que la performance d’un système en se basant sur l’utilité de ses agents, et proposons une méthode pour qu’ils puissent analyser l’évolution de cette performance / With the massive rise of wireless technologies, the number of mobile stations is constantly growing. Both their uses and their communication resources are diversified. By integrating our daily life objects, our communication networks become dynamic in terms of physical topology but also in term of resources. Furthermore, they give access to a richer information. As a result, the management task has become complex and requires shorter response time that a human administrator can not respect. It becomes necessary to develop an autonomic management behavior in next generation networks. In any manner, managing a system requires essential steps which are : its measurement and its supervision. Whatever the nature of a system, this stage of information gathering, allows its characterization and its control. The field of networks is not the exception to the rule and objects that compose them will need to acquire information on their environment for a better adaptation. In this thesis, we focus on the efficient sharing of this information, for self-analysis and distributed performance evaluation purposes. After having formalized the problem of the distributed measurement, we address in a first part the fusion and the diffusion of measures in dynamic graphs. We develop a new heuristic for the average consensus problem offering a better contraction rate than the ones of the state of the art. In a second part, we consider more stable topologies where TCP is used to convey measures. We offer a scheduling mechanism for TCP flows that guaranty the same impact on the network congestion, while reducing the average latency. Finally, we show how nodes can supervise various metrics such as the system performance based on their utilities and suggest a method to allow them to analyze the evolution of this performance

<p>The advent of advanced communication technologies has given rise to large-scale networks comprised of numerous interconnected agents, which need to cooperate to accomplish various tasks, such as distributed message routing, formation control, robust statistical inference, and spectrum access coordination. These tasks can be formulated as distributed optimization problems, which require agents to agree on a parameter minimizing the average of their local cost functions by communicating only with their neighbors. However, distributed optimization algorithms are typically susceptible to malicious (or "Byzantine") agents that do not follow the algorithm. This thesis offers analysis and algorithms for such scenarios. As the malicious agent's function can be modeled as an unknown function with some fundamental properties, we begin in the first two parts by analyzing the region containing the potential minimizers of a sum of functions. Specifically, we explicitly characterize the boundary of this region for the sum of two unknown functions with certain properties. In the third part, we develop resilient algorithms that allow correctly functioning agents to converge to a region containing the true minimizer under the assumption of convex functions of each regular agent. Finally, we present a general algorithmic framework that includes most state-of-the-art resilient algorithms. Under the strongly convex assumption, we derive a geometric rate of convergence of all regular agents to a ball around the optimal solution (whose size we characterize) for some algorithms within the framework.</p>

[pt] Redes de sensores sem fio (WSN) são uma importante alternativa na coleta de dados em diversas situações, tais como no monitoramento de grandes áreas ou áreas que apresentem perigo. A recuperação de dados de WSNs é uma importante atividade que pode obter melhores resultados com o uso de veículos aéreos não tripulados (UAV) como, por exemplo, em relação ao aumento da quantidade de dados coletados e diminuição do tempo entre a coleta dos dados e seu uso. Em particular, áreas tomadas por desastres podem ficar sem recursos de comunicação e com grande risco residual para humanos, momento no qual uma WSN pode ser rapidamente lançada por via aérea e atuar na coleta de dados relevantes até que medidas pertinentes e dedicadas possam ser colocadas em ação. Estudos apresentam abordagens no uso de UAVs para coleta dos dados de WSN, focando principalmente na otimização do caminho a ser percorrido por um único UAV e se baseando em uma comunicação de longo alcance sempre disponível, não explorando a possibilidade da utilização de diversos UAVs ou à limitação do alcance da comunicação. Neste trabalho apresentamos o DADCA, uma abordagem distribuída escalável capaz de coordenadar grupos de UAVs na coleta de dados de WSN sob restrições de alcance de comunicação, sem fazer uso de técnicas de otimização. Resultados indicam que a quantidade de dados coletados pelo DADCA é semelhante ou superior, em até 1 porcento, a abordagens de otimização de caminhos percorridos por UAVs. O atraso no recebimento de mensagens de sensores é até 46 porcento menor do que outras abordagens e o processamento necessário a bordo de UAVs é no mínimo menor do que 75 porcento do que aqueles que utilizam algoritmos baseados em otimização. Os resultados apresentados indicam que o DADCA é capaz de igualar e até superar outras abordagens apresentadas, agregando vantagens de uma abordagem distribuída. / [en] Wireless sensor networks (WSNs) are an important means of collecting data in a variety of situations, such as monitoring large or hazardous areas. The retrieval of WSN data can yield better results with the use of unmanned aerial vehicles (UAVs), for example, concerning the increase in the amount of collected data and decrease in the time between the collection and use of the data. In particular, disaster areas may be left without communication resources and with great residual risk to humans, at which point a WSN can be quickly launched by air to collect relevant data until other measures can be put in place. Some studies present approaches to the use of UAVs for the collection of WSN data, focusing mainly on optimizing the path to be covered by a single UAV and relying on long-range communication that is always available; these studies do not explore the possibility of using several UAVs or the limitations on the range of communication. This work describes DADCA, a distributed scalable approach capable of coordinating groups of UAVs in WSN data collection with restricted communication range and without the use of optimization techniques. The results show that the amount of data collected by DADCA is similar or superior, by up to 1 percent, to path optimization approaches. In the proposed approach, the delay in receiving sensor messages is up to 46 percent shorter than in other approaches, and the required processing onboard UAVs can reach less than 75 percent of those using optimization-based algorithms. The results indicate that the DADCA can match and even surpass other approaches presented, while also adding the advantages of a distributed approach.

Computing with mobile agents is rapidly becoming a topic of mainstream research in the theory of distributed computing. The main research questions undertaken in this study concern the feasibility of solving fundamental tasks in an anonymous network, subject to limitations on the resources available to the agent. The considered challenges include: exploring a graph by means of an agent with limited memory, discovery of the network topology, and attempting to meet with another agent in another network (rendezvous). The constraints imposed on the agent include the number of moves which the agent is allowed to perform in the network, the amount of state memory available to the agent, the ability of the agent to communicate with other agents, as well as its a priori knowledge of the network topology or of global parameters.

Dans un ordinateur multiprocesseur, lors de l'accès à la mémoire partagée, il faut synchroniser les entités de calcul (processus). Cela peut se faire à l'aide de verrous, mais des problèmes se posent (par exemple interblocages, mauvaise tolérance aux pannes). On s'est intéressé à l'implémentation d'abstractions (consensus et construction universelle) qui peuvent faciliter la programmation concurrente sans attente, sans utiliser de verrous mais basés sur des lectures/écritures atomiques (LEA). L'usage exclusive des LEA ne permet pas de réaliser un consensus sans attente. Néanmoins, autoriser l'usage de primitives offrant une puissance de synchronisation plus forte que des LEA, mais coûteuse en temps de calcul, le permet. Nous nous sommes donc intéressés dans cette thèse à des programmes qui limitent l'usage de ces primitives aux seules situations où les processus sont en concurrence, ces programmes sont dit solo-rapides. Une autre piste étudiée est de permettre à l'objet, lorsqu'il y a de la concurrence, de retourner une réponse spéciale "abandon" qui signifie l'abandon des calculs en cours. Ces objets sont dit abandonnables. D'une part, nous donnons des implémentations d'objets concurrents sans attente, abandonnables et/ou solo-rapides. Pour cela, nous proposons une construction universelle qui assure à l'objet implémenté d'être abandonnable et solo-rapide ; nous avons réalisés des algorithmes de consensus solo-rapides et des algorithmes de consensus abandonnable. D'autre part nous étudions la complexité en espace de ces implémentations en proposant des bornes inférieures sur l'implémentation des objets abandonnables et sur le consensus. / In multiprocessor computer, synchronizations between processes are needed for the access to the shared memory. Usually this is done by using locks, but there are some issues as deadlocks or lack of fault-tolerance. We are interested in implementing abstractions (as consensus or universal construction) which ease the programming of wait-free concurrent objects, without using lock but based on atomic Read/Write operations (ARW). Only using the ARW does not permit to implement wait-free consensus. The use of primitives which offer a higher power of synchronization than the ARW is needed. But these primitives are more expensive in computing time. Therefore, we are interested in this thesis in the design of algorithms which restrict the use of these primitives only to the cases where processes are in contention. These algorithms are said solo-fast. Another direction is to allow the object to abort the computation in progress - and to return a special response "abort" - when there is contention. These objects are named abortable. On the one hand we give wait-free, abortable and/or solo-fast concurrent object implementations. Indeed we proposed a universal construction which ensure to the implemented object to be abortable and solo-fast. We have also realized solo-fast consensus algorithms and abortable consensus algorithms. On the other hand, we study the space complexity of these implementations : we prove space lower bound on the implementation of abortable object and consensus.

Recently, a lot of research work has been dedicated toward enhancing performance, reliability and integrity of distributed energy resources that are integrated into distribution networks. The problem of islanding detection and islanding prevention (i.e. anti-islanding) has stimulated a lot of research due to its role in severely compromising the safety of working personnel and resulting in equipment damages. Various Islanding Detection Methods (IDMs) have been developed within the last ten years in anticipation of the tremendous increase in the penetration of Distributed Generation (DG) in distribution system. This work proposes new IDMs that rely on transient and distributed behaviors to improve integrity and performance of DGs while maintaining multi-DG islanding detection capability. In this thesis, the following questions have been addressed: How to utilize the transient behavior arising from an islanding condition to improve detectability and robust performance of IDMs in a distributive manner? How to reduce the negative stability impact of the well-known Sandia Frequency Shift (SFS) IDM while maintaining its islanding detection capability? How to incorporate the perturbations provided by each of DGs in such a way that the negative interference of different IDMs is minimized without the need of any type of communication among the different DGs? It is shown that the proposed techniques are local, scalable and robust against different loading conditions and topology changes. Also, the proposed techniques can successfully distinguish an islanding condition from other disturbances that may occur in power system networks. This work improves the efficiency, reliability and safety of integrated DGs, which presents a necessary advance toward making electric power grids a smart grid.