O gerenciamento de redes é realizado seguindo diversos modelos, baseados em diferentes paradigmas. Os paradigmas tradicionais de gerenciamento compreendem o paradigma centralizado e o paradigma hierárquico fracamente distribuído. Tais paradigmas, contudo, apresentam limitações para o gerenciamento de diversas redes atuais, em virtude de fatores como o aumento em tamanho, em complexidade e em heterogeneidade destas redes. Em adição a estes fatores, existem atualmente contextos de rede que, por apresentarem certas peculiaridades, não podem ser gerenciados de modo apropriado por tais modelos. Estas limitações e requisitos diferenciados encontrados nas redes atuais trazem a necessidade do emprego de modelos de gerenciamento inovadores, baseados nos paradigmas fortemente distribuídos. Neste contexto, uma tecnologia que se mostra promissora para o desenvolvimento de um modelo de gerenciamento com tais características é a tecnologia peer-to-peer (P2P). Esta Tese versa sobre estes tópicos. Discute, como primeiro objetivo, as limitações dos modelos tradicionais para o gerenciamento de contextos modernos de rede, isto é, contextos de redes atuais que possuem particularidades distintas e, com isto, demandam requisitos de gerenciamento específicos, não identificados nas redes tradicionais. Define e investiga, então, como segundo objetivo, um modelo de gerenciamento fortemente distribuído baseado na tecnologia P2P. Este modelo visualiza a rede P2P como uma infra-estrutura que provê suporte para que as operações de gerenciamento sejam desempenhadas com forte distribuição. Por fim, como terceiro objetivo, analisa como as soluções de gerenciamento para redes atuais baseadas neste modelo podem ser classificadas segundo os paradigmas de gerenciamento das principais taxonomias propostas na literatura. Tal análise proporciona a identificação das limitações destas taxonomias para a classificação das soluções de gerenciamento requeridas para os contextos de redes atuais e deu origem à definição de uma taxonomia para soluções de gerenciamento que destaca as características e os requisitos demandados dos modelos de gerenciamento modernos. / Network management is carried out following several models, based on different paradigms. Traditional management paradigms consist of centralized and weakly distributed hierarchical ones. However, such paradigms present limitations to be applied to the management of several today networks. This occurs because of some issues such as the grown in size, complexity and heterogeneity of such networks. Additionally, nowadays, there are network contexts that can not be appropriately managed by such models because of some context peculiarities. Those today’s networks drawbacks and requirements demand the employment of innovative models, based on strongly distributed paradigms. A technology that seems promising in addressing such needs is peer-to-peer (P2P). This Thesis discusses about those topics. As its first objective, the Thesis discusses traditional models drawbacks to the management of modern network contexts, this is, current network contexts that have some different peculiarities and, because of them, demand specific management requirements not existent in traditional networks. As its second objective, the Thesis defines and investigates a strongly distributed management model based on P2P technology. Such model looks at P2P network as an infrastructure that can be used as support to management operations be accomplished in a strongly distributed way. Finally, as its third objective, the Thesis analyses how the management solutions based on such model can be classified according the management paradigms of the main literature taxonomies. Such analysis provides the identification of taxonomies limitations to the classification of management solutions required by today’s network contexts. It has originated the definition of a management solution taxonomy that emphasizes the features and requirements demanded of modern management models.

Com o crescimento em número e diversidade dos componentes das redes de computadores, surge a necessidade de buscar uma maneira consistente de realizar seu gerenciamento para, com isso, manter toda sua estrutura funcionando de forma suave e atendendo às necessidades de seus usuários e às expectativas de seus administradores. Em aspectos gerais, as abordagens centralizadas têm se mostrado inadequadas para o gerenciamento de redes de computadores com um grande número de nós ou com grande diversidade de dispositivos. A necessidade de distribuição da gerência torna-se, assim, evidente, onde vários operadores administram cooperativamente a rede, cada um sendo responsável por um segmento da mesma, mas que precisam poder inspecionar os demais segmentos para poder resolver problemas distintos. Por isso a necessidade de um sistema de suporte ao gerenciamento cooperativo, que forneça suporte à interação dos administradores, independente de tempo e da localização dos participantes. É esse sistema que determinará como os administradores vão se comunicar, distribuir responsabilidades, compartilhar informações, e utilizar as ferramentas disponíveis. Nesse contexto, esta dissertação de mestrando apresenta a proposta de um ambiente de gerenciamento distribuído e cooperativo, baseado na tecnologia P2P. Esse ambiente oferece quatro serviços: compartilhamento de arquivos de configuração de dispositivos, compartilhamento de registros de dispositivos, configuração de dispositivos e solicitação de reserva de banda. A partir da proposta do ambiente foi implementada em Java, utilizando a plataforma de desenvolvimento JXTA, uma aplicação para a realização de testes, com o objetivo de confirmar a possibilidade de utilização da aplicação para realizar o gerenciamento de redes. Os parâmetros observados nas avaliações de desempenho foram o tráfego gerado e o tempo de resposta. Os resultados dos testes comprovaram a possibilidade da utilização de sistemas P2P para facilitar o gerenciamento cooperativo de redes e foram bastante satisfatórios com relação aos parâmetros avaliados. / With the increasing number and diversity of components of the computer networks, comes the necessity to find a consistent way to manage modern networks. This is required in order to keep all network infrastructures working in a proper and smooth way, while taking care of to the network users’ necessities and filling the network administrators’ expectations. In general aspects, the centralized management approaches are inadequate for the management of networks with a great number of nodes or great diversity of devices. The necessity of a distributed management thus becomes evident. In such distributed management, a group of administrators manages a single network in a cooperative fashion, each administrator being responsible for a segment of the managed network but at the same time being able to inspect other remote segments in order to solve non-overlapping problems. That is the motivation for having management systems with proper cooperative management support that allows interactions among administrators independent of time and location. Such a system will determine how administrators will communicate to delegate responsibilities, share information, and use the available tools. In this context, this dissertation presents the proposal of a distributed and cooperative management environment based on P2P technology. This environment offers four services: sharing of device configuration files, sharing of registered devices, configuration of devices, and bandwidth reservation. Based on the proposed environment, a Java application, using the JXTA development platform, has been implemented. This application allowed us to test and confirm the possibility of using P2P technologies for network management. In addition, the performance of the implemented solution has been measured considering the generated management traffic and response time. The results from the evaluation tests had proven the possibility of the use of P2P systems to facilitate the cooperative network management and had been sufficiently satisfactory in relation to the evaluated parameters.

In the past decade or so, significant advances were made in the field of Future Internet Architecture (FIA) design. Undoubtedly, the size of Future Internet will increase tremendously, and so will the complexity of its users’ behaviors. This advancement means most of future Internet applications and services can only achieve and demonstrate full potential on a large-scale basis. The development of network testbeds that can validate key design decisions and expose operational issues at scale is essential to FIA research. In conjunction with the development and advancement of FIA, cyber-infrastructure testbeds have also achieved remarkable progress. For meaningful network studies, it is indispensable to utilize cyber-infrastructure testbeds appropriately in order to obtain accurate experiment results. That said, existing current network experimentation is intrinsically deficient. The existing testbeds do not offer scalability, flexibility, and realism at the same time. This dissertation aims to construct a hybrid system of conducting at-scale network studies and experiments by exploiting the distributed computing ability of current testbeds. First, this work presents a synchronization of parallel discrete event simulation that offers the simulation with transparent scalability and performance on various high-end computing platforms. The parallel simulator that we implement is configured so that it can self-adapt for the performance while running on supercomputers with disparate architectures. The simulator could be used to handle models of different sizes, varying modeling details, and different complexity levels. Second, this works addresses the issue of researching network design and implementation realistically at scale, through the use of distributed cyber-infrastructure testbeds. An existing symbiotic approach is applied to integrate emulation with simulation so that they can overcome the limitations of physical setup. The symbiotic method is used to improve the capabilities of a specific emulator, Mininet. In this case, Mininet can be used to run applications directly on the virtual machines and software switches, with network connectivity represented by detailed simulation at scale. We also propose a method for using the symbiotic approach to coordinate separate Mininet instances, each representing a different set of the overlapping network flows. This approach provides a significant improvement to the scalability of the network experiments.

Les systèmes collaboratifs sont aujourd’hui devenus très populaires et sont de plus en plus utilisés dans de nombreux domaines divers. De fait, les interactions de confiance des differents systèmes sont devenus une priorité. La confiance, en tant que concept informatique, a été étudiée très récemment. Cependant, dans la littérature, très peu d’attention a été portée pour évaluer l’exactitude des interactions entre entités communicantes; même si la plupart des approches se sont basées sur les mesures cumulées de ces valeurs. Pour déterminer, de façon générale, l’exactitude de ces interactions, une approche nommée Monitorage des Réseaux En-Ligne et Distribué (MRED) a été proposée. De plus, des outils prototypes ont été développés pour tester automatiquement les propriétés de confiance entre entités dans des systèmes communicants. MRED est une forme de test passif; elle analyse les réponses des systèmes et teste l’exactitude des interactions en utilisant des traces de réseaux. Comme elle dépend des propriétés à tester, une nouvelle approche a été proposée pour faire l’extraction automatique de propriétés pertinentes que l’ont pourrait, in fine, tester dans un système sous test. Notre approche repose sur le fonctionnement des systèmes de monitorage online. Nous proposons de nouvelles méthodes afin d’améliorer les techniques fournies dans l’état de l’art pour: a) évaluer efficacement les propriétés avec une complexité en temps O($n$), ce en utilisant un Automate Fini Déterministe Prolongée (AFDP); et b) élargir l’expressivité du langage proposé pour exprimer correctement les contraintes systèmes, comme les délais d’attente pour éviter le manque de ressources. Finalement, nous proposons un nouveau cadre flexible utilisable dans de très nombreux domaines, qui permet la définition de caractéristiques de confiance afin d’évaluer les entités dans des contextes différents. De surcroît, avec les évaluations des caractéristiques de confiance, nous proposons un modèle de confiance basé sur l’apprentissage automatique, en résolvant spécifiquement un problème de classification multi-classes et utilisant des Machine à vecteurs de support (SVM). A partir de ces modèles, des expérimentations ont été effectuées en simulant des caractéristiques de confiance pour estimer le niveau de confiance; une précision de plus de 96% a été obtenue / Collaborative systems are growing in use and in popularity. The need to boost the methods concerning the interoperability is growing as well; therefore, trustworthy interactions of the different systems are a priority. Trust as a computer science concept has been studied in the recent years. Nevertheless, in the literature, very little focus is placed on how to assess the correctness of the interactions between the entities; even if most approaches rely on the estimation of trust based on the accumulated measures of these values. To broadly determine the correctness of interactions without targeting a specific domain or application, an approach using Distributed On-line Network Monitoring (DONM) was proposed. Furthermore, a prototype tool-set was developed to automatically test the trust properties. DONM is a form passive testing; it analyzes systems' responses and test the correctness of the interactions via network traces. Since it relies on the stated properties to test, a novel approach was proposed to automatically extract relevant properties to test. Our approach deeply relies on the operation of On-line Monitoring Systems. That is the reason why we propose new methods to enhance the state of the art techniques to: a) efficiently evaluate properties in O(n) time complexity using an Extended Finite State Automata (EFSA) auxiliary data structure; and b) to expand the language expressiveness to properly express the constraints of such systems, such as, timeouts in order to avoid resource starvation. Finally, using the evaluation of the entities' interactions provided by our approach, trust management engines will help trustors to decide with whom and how to interact with other users or applications. We propose a new framework that is flexible for any domain, allowing trustors to define the trust features used to evaluate trustees in different contexts. Furthermore, with the evaluations of the trust features, we propose a trust model which achieves close-to-human inference of the trust assessment, by using a machine learning based trust model, namely solving a multi-class classification problem using Support Vector Machines (SVM). Using the SVM-based trust model, experiments were performed with simulated trust features to estimate trust level; an accuracy of more than 96% was achieved

This thesis contributes to the development of distributed networks proposing: • a technique to enhance the reliability of DV routing protocols; • a critical analysis of the integration of blockchains in distributed networks. First, a novel algorithm for the distributed computation of the Load Centrality (LC), a graph centrality metric, is proposed and then applied for steering the optimization of the route recovery process of Distance-Vector (DV) routing protocols: this way the algorithm contributes to the enhancement of the network reliability. The algorithm convergence is proved also identifying time complexity bounds that are later confirmed by computer simulations. The proposed algorithm is designed as an extension to the Bellman-Ford one and can thus be integrated with any DV routing protocol. An implementation of the algorithm in Babel, a real world DV protocol, is provided in support of this claim. Then an application of the algorithm is presented: the LC is used to find an optimal tuning for the generation frequency of the Babel control messages. This tuning technique effectively reduces the impact of losses consequent to random node failures in the emulations of several real world wireless mesh networks, without increasing the control overhead. A second version of the algorithm is designed to be incrementally deployable. This version can be deployed gradually in production networks also by uncoordinated administrators. When only a fraction of nodes is upgraded so to participate in the protocol, these upgraded nodes estimate their LC indexes approximating the theoretical ones. The approximation error is studied analytically and it is also shown that, even for low penetration ratios of upgraded nodes in the network, the algorithm accurately ranks nodes according to their theoretical centrality. The second contribution of the thesis is the critical discussion of the integration of blockchain technologies in distributed networks. An initial analysis of the literature concerning blockchain based applications reveals an ambiguity around the term "blockchain" itself. The term is used, apparently, to identify a number of similar but different technologies proposed to empower a surprisingly broad range of applications. This thesis prompts therefore the need of formulating a restrictive definition for the term blockchain, necessary for clarifying the role of the same blockchain in distributed networks. The proposed definition is grounded in the critical analysis of the blockchain from a distributed systems perspective: Blockchains are only those platforms that implement an open, verifiable and immutable Shared Ledger, independent of any trusted authority. Observing that the blockchain security grows with the amount of resources consumed to generate blocks, this thesis concludes that a secure blockchain is necessarily resource hungry, therefore, its integration in the constrained domain of distributed networks is not advised. The thesis draws recommendations for a use of the blockchain not in contrast with the definition. For example, it warns about applications that require data to be kept confidential or users to be registered, because the blockchain naturally supports the openness and transparency of data together with the anonymity of users. Finally a feasible role for the blockchain in the Internet of Things (IoT) is outlined: while most of the IoT transactions will be local and Off-Chain, a blockchain can still act as an external and decentralized platform supporting global transactions, offering an alternative to traditional banking services. The enhanced reliability of DV routing protocols encourages a wider adoption of distributed networks, moreover, the distributed algorithm for the computation of centrality enables applications previously restricted to centralized networks also in distributed ones. The discussion about the blockchain increases instead the awareness about the limits and the scope of this technology, inspiring engineers and practitioners in the development of more secure applications for distributed networks. This discussion highlights, for instance, the important role of the networking protocols and communication infrastructure on the blockchain security, pointing out that large delays in the dissemination of blocks of transactions make the blockchain more vulnerable to attacks. Furthermore, it is observed that a high ability to take control over the communications in the network favors eclipse attacks and makes more profitable the so called selfish mining strategy, which is detrimental to the decentralization and the security of blockchains. The two main contributions of this thesis blended together inspire the exploitation of centrality to optimize gossip protocols, minimizing block propagation delays and thus the exposure of the blockchain to attacks. Furthermore, the notion of centrality may be used by the community of miners to measure the nodes influence over the communication of blocks, so it might be used as a security index to warn against selfish mining and eclipse attack.

Cette thèse adresse le problème du consensus dans les réseaux. On étudie des réseaux composés d'agents identiques capables de communiquer entre eux, qui ont une mémoire et des capacités de calcul. Le réseau ne possède pas de nœud central de fusion. Chaque agent stocke une valeur qui n'est pas initialement connue par les autres agents. L'objectif est d'atteindre le consensus, i.e. tous les agents ont la même valeur, d'une manière distribuée. De plus, seul les agents voisins peuvent communiquer entre eux. Ce problème a une longue et riche histoire. Si toutes les valeurs appartiennent à un espace vectoriel, il existe plusieurs protocoles pour résoudre le problème. Une des solutions connues est l'algorithme du gossip qui atteint le consensus de manière asymptotique. C'est un protocole itératif qui consiste à choisir deux nœuds adjacents à chaque itération et de les moyenner. La spécificité de cette thèse est dans le fait que les données stockées par les agents n'appartiennent pas nécessairement à un espace vectoriel, mais à un espace métrique. Par exemple, chaque agent stocke une direction (l'espace métrique est l'espace projectif) ou une position dans un graphe métrique (l'espace métrique est le graphe sous-jacent). Là, les protocoles de gossip mentionnés plus haut n'ont plus de sens car l'addition qui n'est plus disponibles dans les espaces métriques. Cependant, dans les espaces métriques les points milieu ont du sens dans certains cas. Et là ils peuvent se substituer aux moyennes arithmétiques. Dans ce travail, on a compris que la convergence du gossip avec les points milieu dépend de la courbure. On s'est focalisés sur le cas où l'espace des données appartient à une classe d'espaces métriques appelés les espaces CAT(k). Et on a pu démontrer que si les données initiales sont suffisamment "proches" dans un sens bien précis, alors le gossip avec les points milieu - qu'on a appelé le Random Parwise Midpoints- converge asymptotiquement vers un consensus / This thesis deals with the problem of consensus on networks. Networks under study consists of identical agents that can communicate with each other, have memory and computational capacity. The network has no central node. Each agent stores a value that, initially, is not known by other agents. The goal is to achieve consensus, i.e. all agents having the same value, in a fully distributed way. Hence, only neighboring agents can have direct communication. This problem has a long and fruitful history. If all values belong to some vector space, several protocols are known to solve this problem. A well-known solution is the pairwise gossip protocol that achieves consensus asymptotically. It is an iterative protocol that consists in choosing two adjacent nodes at each iteration and average them. The specificity of this Ph.D. thesis lies in the fact that the data stored by the agents does not necessarily belong to a vector space, but some metric space. For instance, each agent stores a direction (the metric space is the projective space) or position on a sphere (the metric space is a sphere) or even a position on a metric graph (the metric space is the underlying graph). Then the mentioned pairwise gossip protocols makes no sense since averaging implies additions and multiplications that are not available in metric spaces: what is the average of two directions, for instance? However, in metric spaces midpoints sometimes make sense and when they do, they can advantageously replace averages. In this work, we realized that, if one wants midpoints to converge, curvature matters. We focused on the case where the data space belongs to some special class of metric spaces called CAT(k) spaces. And we were able to show that, provided initial data is "close enough" is some precise meaning, midpoints-based gossip algorithm – that we refer to as Random Pairwise Midpoints - does converge to consensus asymptotically. Our generalization allows to treat new cases of data spaces such as positive definite matrices, the rotations group and metamorphic systems

In this research, the application of intelligent mobile agents to the management of distributed network environments is investigated. Intelligent mobile agents are programs which can move about network systems in a deterministic manner in carrying their execution state. These agents can be considered an application of distributed artificial intelligence where the (usually small) agent code is moved to the data and executed locally. The mobile agent paradigm offers potential advantages over many conventional mechanisms which move (often large) data to the code, thereby wasting available network bandwidth. The performance of agents in network routing and knowledge acquisition has been investigated and simulated. A working mobile agent system has also been designed and implemented in JDK 1.2.

This thesis report, focuses on the automation process of Electrical Substations, with the main objective of integration into SCADA systems for electrical equipment responsible for the operation within substations, these devices have different communication protocols, after concentration information substation SCADA system, the next step to realize is sending data to the regulatory body that is in a remote location away from the substation, this is achieved using communication protocols designed for this function. All this process requires an optimization for the response times to any event that may happen in the substations and with this an optimization in operation of generation, distribution and transformation of electrical energy

Diversification of the energy mix is one of the main challenges in the energy agenda of governments worldwide. Technology advances together with environmental concerns have paved the way for the increasing integration of Distributed Generation (DG) seen over recent years. Combined heat and power and renewable technologies are being encouraged and their penetration in distribution networks is increasing. This scenario presents Distribution Network Operators (DNOs) with several technical challenges in order to properly accommodate DG developments. However, depending on various factors, such as location, size, technology and robustness of the network, DG might also be beneficial to DNOs. In this thesis, the impact of DG on network planning is analysed and the implications for DNOs in incorporating DG within the network planning process are identified. In the first part, various impacts of DG to the network, such as network thermal capacity release, security of supply and on voltage, are quantified through network planning by using a modified successive elimination method and voltage sensitivity analysis. The results would potentially assist DNOs in assessing the possibilities and effort required to utilise privately-owned DG to improve network efficiency and save investment. The quantified values would also act as a fundamental element in deriving effective distribution network charging schemes. In the second part, a novel balanced genetic algorithm is introduced as an efficient means of tackling the problem of optimum network planning considering future uncertainties. The approach is used to analyse the possibilities, potential benefits and challenges to strategic network planning by considering the presence of DG in the future when the characteristics of DG are uncertain.

Cette thèse porte sur la notion de certification locale, un sujet central en décision distribuée, un domaine du calcul distribué. Le mécanisme de la décision distribuée consiste, pour les nœuds d'un réseau, à décider de manière distribuée si le réseau est dans une configuration correcte ou non, selon un certain prédicat. Cette décision est dite locale, car les nœuds du réseau ne peuvent communiquer qu'avec leurs voisins. Après avoir communiqué, chaque nœud prend une décision, exprimant si le réseau est correct ou non localement, c'est-à-dire correct étant donné l'information partielle récoltée jusque-là. Le réseau est déclaré correct globalement s'il est déclaré correct localement par tous les nœuds.Du fait de la contrainte de localité, peu de prédicats peuvent être vérifiés de cette manière. La certification locale est un moyen de contourner cette difficulté, et permet de décider tous les prédicats. C'est un mécanisme qui consiste à étiqueter les nœuds du réseau avec ce que l'on appelle des certificats, qui peuvent être vérifiés localement par un algorithme distribué. Un schéma de certification locale est correct si seuls les réseaux dans une configuration correcte peuvent être certifiés. L'idée de la certification locale est non seulement séduisante d'un point de vue théorique, comme une forme de non-déterminisme distribué, mais c'est surtout un concept très utile pour l'étude des algorithmes tolérants aux pannes, où une étape-clé consiste à vérifier l'état du réseau en se basant sur des informations stockées par les nœuds.Cette thèse porte sur quatre aspects de la certification locale : la sensibilité aux erreurs, l'uniformité, la redondance et l'interactivité. L'étude de ces quatre sujets est motivée par une question essentielle : comment réduire les ressources nécessaires à la certification et/ou permettre une meilleure tolérance aux pannes? Pour aborder cette question, il est nécessaire de comprendre le mécanisme de certification en profondeur. Dans cette optique, dans cette thèse, nous apportons des réponses aux questions suivantes. À quel point les certificats doivent-ils être redondants, pour assurer une certification correcte? Les schémas de certification classiques sont-ils robustes à un changement de la condition de correction? Le fait d'introduire de l'interactivité dans le processus change-t-il la complexité de la certification? / This dissertation is about local certification, a central topic in distributed decision, a subfield of distributed computing. The distributed decision mechanism consists, for the nodes of a network, in deciding in a distributed manner whether the network is in a proper configuration or not, with respect to some fixed predicate. This decision is said to be local because the nodes of the network can communicate only with their neighbours. After communication, every node outputs a decision, stating whether the network is locally correct, that is, correct given the partial information gathered so far by this node. The network is declared to be globally correct, if and only if, it is declared to be locally correct by every node.Most predicates cannot be verified by this type of computation, due to the locality constraint. Local certification is a mechanism that enables to circumvent this difficulty, and to check any property. It consists in providing the nodes of the network with labels, called certificates, that can be verified locally by a distributed algorithm. A local certification scheme is correct if only the networks that satisfy the predicate can be certified. In addition to its theoretical appeal, as a form of distributed non-determinism, the concept of local certification is especially relevant in the study of fault-tolerant distributed algorithms, where a key step consists in checking the status of the network, based on information stored at the nodes.This dissertation deals with four aspects of local certification: error-sensitivity, uniformity, redundancy, and interactivity. The study of these four topics is motivated by the same essential question: How to reduce the resources needed for certification, and/or ensure a better fault-tolerance? In order to tackle this question we have to understand fundamental properties of certification. In particular, in this dissertation we answer questions such as: How redundant the certificates need to be for a proper certification? Are the classic certification protocols robust to a strengthening of the acceptance condition? and, How does introducing interactivity in the process changes the complexity of certification?