Τα δίκτυα αισθητήρων είναι μια εξειδικευμένη κατηγορία κατανεμημένων δικτύων, η οποία τα τελευταία χρόνια έχει συγκεντρώσει το ενδιαφέρον της ερευνητικής κοινότητας, λόγω του ευρύτατου πεδίου εφαρμογών της. Τα δίκτυα αυτά αποτελούνται από συσκευές που διαθέτουν αισθητήρες (sensors) και ενδεχομένως μηχανισμούς δράσης (actuators) και είναι διασκορπισμένες στο χώρο, με δυνατότητα επικοινωνίας μεταξύ τους και επεξεργασίας σε ένα βαθμό της πληροφορίας που διακινούν στο δίκτυο. Στόχος των δικτύων αυτών είναι η παρακολούθηση της εξέλιξης ενός φαινόμενου, ή η ανίχνευση περιβαλλοντικών συνθηκών και η αποστολή των δεδομένων που συλλέγονται σε έναν κεντρικό κόμβο, το κέντρο ελέγχου. Οι δυνατότητες των συσκευών είναι κατά κανόνα περιορισμένες, λόγω του μικρού μεγέθους τους, του χαμηλού κόστους τους και ενίοτε του αναλώσιμου ρόλου τους. Οι εφαρμογές των δικτύων αυτών ποικίλουν, από εφαρμογές στην επιστήμη γεωργίας ακριβείας, στην πυρανίχνευση και την παρακολούθηση συνθηκών περιβάλλοντος σε κτιριακές εγκαταστάσεις, μέχρι εφαρμογές ανίχνευσης κίνησης των εχθρικών μονάδων σε πεδίο μάχης, ή παρακολούθησης φαινόμενου που εξελίσσεται σε δυσπρόσιτη περιοχή. Μερικές από τις βασικές προκλήσεις που προκύπτουν σε αυτά τα δίκτυα, λόγω της φύσης τους, είναι η βέλτιστη διαχείριση των ενεργειακών πόρων κάθε κόμβου, η αποδοτική συνδυαστική επεξεργασία των διακινούμενων μηνυμάτων και η αξιόπιστη δρομολόγηση της πληροφορίας που συλλέγεται προς το κέντρο ελέγχου, καθώς και η δυναμική και αυτόνομη οργάνωση του δικτύου. Η εξέλιξη των δικτύων αισθητήρων σε συνδυασμό με την επιθυμία υλοποίησης περίπλοκων και περισσότερο ολοκληρωμένων εφαρμογών, οδήγησε σε δίκτυα που περιέχουν ετερογενείς κόμβους, όπως για παράδειγμα κόμβους με διαφορετική αρχιτεκτονική, διαφορετικούς αισθητήρες, με διαφορετικές επεξεργαστικές και επικοινωνιακές ικανότητες ή και κόμβους με δυνατότητες κίνησης ή επιτέλεσης εργασιών (actuators) σε απόκριση εντολών που λαμβάνουν από το κέντρο ελέγχου. Επιπρόσθετα, σε πολλές εφαρμογές διαφαίνεται η ανάγκη για διασύνδεση πολλαπλών δικτύων αισθητήρων και διάδοση της πληροφορίας που συλλέγουν μέσω του διαδικτύου προς απομακρυσμένους εξυπηρετητές με αυξημένες επεξεργαστικές ικανότητες, προς βάσεις δεδομένων, συσκευές pda ή κινητά τηλέφωνα, σταθμούς εργασίας τελικών χρηστών και άλλες συσκευές. Δημιουργείται με αυτόν τον τρόπο ένα ευρύτερο ετερογενές δίκτυο επικάλυψης (overlay sensor network), η διαχείριση και αποδοτική χρήση του οποίου απαιτεί νέες αρχιτεκτονικές οργάνωσης των επιμέρους δικτύων, προσαρμοσμένα πρωτόκολλα διάδοσης πληροφορίας και πλατφόρμες που διευκολύνουν τον προγραμματισμό των δικτύων αυτών. Για το λόγο αυτό η επιστημονική κοινότητα επικεντρώνει το ενδιαφέρον της στο σχεδιασμό ειδικών περιβαλλόντων λογισμικού που μεσολαβούν ανάμεσα στον προγραμματιστή εφαρμογών και τα ετερογενή δίκτυα αισθητήρων, τα οποία ανήκουν στην κατηγορία των middleware. Τα περιβάλλοντα αυτά αναλαμβάνουν να διαχειριστούν θέματα χαμηλού επιπέδου ή και ανώτερων επιπέδων στα δίκτυα αισθητήρων, που δεν θα έπρεπε να απασχολούν τον προγραμματιστή κάθε φορά που επιθυμεί να υλοποιήσει μια εφαρμογή. Απώτερος σκοπός είναι να παρέχουν υπηρεσίες υψηλού επιπέδου και ομοιογενείς λογικές αφαιρέσεις πάνω στα ετερογενή δίκτυα, ενώ είναι επιθυμητή η αποδοτική λειτουργία τους σε παγκόσμια κλίμακα όπου πολλά γεωγραφικά διάσπαρτα δίκτυα συνεργάζονται για να παρέχουν τις υπηρεσίες τους με οργανωμένο και καλά ορισμένο τρόπο. Η δική μας συνεισφορά στην έρευνα του πεδίου αυτού, έγκειται πρώτα στη μελέτη ήδη διαθέσιμων προτάσεων middleware για τον προγραμματισμό και τη διασύνδεση πολλαπλών και ετερογενών δικτύων αισθητήρων, με σκοπό τον εντοπισμό των θεμάτων και προβλημάτων που αντιμετώπισαν κατά το σχεδιασμό τους, την ανάλυση της αρχιτεκτονικής τους και την αξιολόγηση της απόδοσής τους. Επιθυμώντας να δώσουμε τη δική μας πρόταση στο πεδίο των middleware για δίκτυα αισθητήρων και αξιοποιώντας τα συμπεράσματά μας από την μελέτη των άλλων προτάσεων, σχεδιάσαμε και υλοποιήσαμε το middleware σύστημα ShareSense II. Το σύστημα ShareSense II προέρχεται από τον επανασχεδιασμό τού αρχικού και απλούστερου ShareSense και ενσωματώνει το middleware jWebDust επεκτείνοντας κατά ένα τρόπο την αρχιτεκτονική του, ώστε να υποστηρίζει δίκτυα επικάλυψης (overlay sensor networks) μεγάλης κλίμακας. Η πλατφόρμα του ShareSense II σχεδιάστηκε για να παρέχει λογική αφαίρεση των δικτύων αισθητήρων, ώστε ο προγραμματιστής να μπορεί εύκολα να ρυθμίσει τις υπηρεσίες που παρέχει ένα δίκτυο επικάλυψης, να διαχειρίζεται τις λειτουργίες των επιμέρους δικτύων, να μπορεί να προγραμματίσει δυναμικά τη συμπεριφορά τους και να καθορίσει τον τρόπο παρουσίασης της χρήσιμης πληροφορίας που αποκομίζει από αυτά, στον τελικό χρήστη. Βασιζόμενοι στην ευέλικτη υποδομή των peer-to-peer δικτύων, προτείνουμε μια λύση για τη διασύνδεση πολλαπλών ετερογενών δικτύων αισθητήρων με συσκευές και σταθμούς εργασίας πάνω από το διαδίκτυο. Δώσαμε έμφαση στην επεκτασιμότητα της αρχιτεκτονικής ως προς τις υποστηριζόμενες υπηρεσίες και τύπους δικτύων, την αξιοποίηση και τη συνεργασία με υπάρχοντα middleware που διαχειρίζονται κάποια επιμέρους δίκτυα και την παροχή χρήσιμων διεπαφών για την εύκολη υλοποίηση πρακτικών εφαρμογών. Υλοποιήσαμε, τέλος, μια εφαρμογή η οποία στηρίζεται στην πλατφόρμα του ShareSense II για να τεκμηριώσουμε την επάρκεια της αρχιτεκτονικής του συστήματός μας και να εντοπίσουμε τα σημεία όπου απαιτείται βελτίωσή της. Η εφαρμογή χρησιμοποιεί το περιβάλλον του Google Earth για την παρακολούθηση των επιμέρους δικτύων που συμμετέχουν στο σύστημα, επιτρέποντας παράλληλα τη δυναμική υποβολή ερωτημάτων και την αντιστοίχιση των δικτύων σε τρισδιάστατα μοντέλα κτιρίων και άλλων περιοχών, ώστε να είναι άμεση και εύληπτη η παρουσίαση της πληροφορίας στον τελικό χρήστη. / Sensor networks are a special case of distributed networks, which in the recent years have become increasingly interesting to the scientific community, due to their wide range of applications. These networks are comprised of devices that have sensors attached to them and possibly even actuators, and are scattered in a field, being able to comminucate with each other and process -to a certain degree- information that they forward in the network. Among the goals of these networks are the monitoring of physical phenomena, the detection of special events or conditions and consequently the communication of the acquired information towards a control center. These devices usually have strict constraints on their capabilities due to their small size, their low production cost and quite often their expendable role. The applications of these networks vary, from applications in the precision agriculture field, fire detection and monitoring environmental conditions in buildings and offices, to applications for the detection of enemy units' movements in the battlefield or observing a physical phenomenon in a inaccessible location. Some of the important challeges that emerge for these networks are the optimal management of the power resources of each node, the efficient aggregation of the in-network messages and the reliable routing of information acquired towards the control center, as well as the autonomous and adaptive network operation. The evolution in the sensor netorks research in combination with the desire to implement even more sophisticated applications, has led to highly heterogeneous networks, where nodes belong to different architectures, they use various sensor models, and they have different processing and communicating capabilities. Additionally, in many applications there is the need to interconnect multiple sensor networks and to distribute the information collected locally towards remote servers, databases, pda devices or cell phones, workstations and other end user devices. Thus, a new type of networks, the overlay networks, emerges. The administration and effecient usage of these networks require new management architectures for the local networs, adaptive protocols for the distribution of information and new software platforms that simplify application programming. Following these developments, the scientific community has focused in designing middleware software for the sensor networks. Such software have the role of managing low level or intermediate level issues, that should not burden a programmer every time s/he needs to design and implement a new application. The main goal of middleware is to provide high level services and unified logical abstractions over heterogeneous sensor networks, and to perform well in a global scale where multiple geographically disparate networks cooperate to provide services in an organized and well defined manner. Our contribution to this field, is foremost to study the state of the art middleware projects for overlay sensor networks, in an effort to understand the issues and problems that were faced during their design, to analyze their architecure and evaluate their performance. We have used the results of this study to design and implement a new middleware architecture, which we have named ShareSense II. Our middleware comes from the redesign and expansion of its simpler predecessor, the original ShareSense middleware, and also integrates the jwebdust middleware, bringing the services of that middleware to the large scal overlay networks. The ShareSense II platfrom was designed to provide logical abstractions for sensor networks, so that the programmer is able to tweak their services, maange the resources of the local networks and build flexible applications. For our overlay netowrk we have embedded an efficient peer-to-peer infrastructure that performs well in networks with frequent and temporary node disconnections and large scale applications. Finally, we implemented an application that uses the ShareSense II middleware, in order to show the benefits of our design and also locate the parts of the architecture that need further optimization. Our application uses the Google Earth software to monitor geographically disparate networks that participate in ShareSense II, allowing the online issuing of queries, as well as matching 3d models of buildings (and other areas) to local networks, in order to display information to the end user in an interesting and exciting way.

O gerenciamento das redes de comunicações modernas frequentemente requer que operadores de diferentes domínios administrativos interajam uns com os outros de modo a realizar uma tarefa de gerenciamento de maneira cooperativa. Nesse contexto, pode-se observar que há distribuição tanto na execução das ações como na tomada das decisões pelos operadores participantes da sessão de gerenciamento. As soluções tradicionais de gerenciamento de redes praticamente não oferecem suporte algum ao gerenciamento baseado na cooperação entre humanos, pois foram concebidas levando em consideração um único domínio administrativo sendo gerenciado por um único operador ou por um grupo de operadores pertencentes a uma mesma organização. Visando solucionar esse problema, é introduzido nesse trabalho o uso de tecnologias peer-to-peer (P2P) em soluções de gerenciamento de redes de modo a permitir o gerenciamento baseado na cooperação entre humanos. Nessa abordagem, operadores de rede usam ferramentas híbridas que, ao mesmo tempo, gerenciam os seus domínios locais e fazem parte de um sistema P2P de gerenciamento. A proposta de gerenciamento apresentada foi implementada através de uma ferramenta chamada ManP2P, a qual combina funcionalidades de gerenciamento de redes e trabalho cooperativo. / Modern networks not rarely require human operators from different administrative domains to interact with one another in order to accomplish a management task in a cooperative fashion. In this scenario one observes that not only management actions are executed in a distributed way, but also the management decisions are cooperatively taken by all autonomous operators participating in a management session. Traditional management solutions fail to support human centered cooperative management because they were conceived considering a single administrative domain being managed by a single operator or by a group of operators belonging to the same organization. In order to address this issue, we introduce in this paper the use of peer-to-peer (P2P) technologies in network management solutions to enable human-centered cooperative management. In this approach, network operators use hybrid tools that both manage their local domains and are peers of an overlay management system. We have implemented the proposed management approach through a tool called ManP2P, which combines network management and groupware functionalities.

O gerenciamento das redes de comunicações modernas frequentemente requer que operadores de diferentes domínios administrativos interajam uns com os outros de modo a realizar uma tarefa de gerenciamento de maneira cooperativa. Nesse contexto, pode-se observar que há distribuição tanto na execução das ações como na tomada das decisões pelos operadores participantes da sessão de gerenciamento. As soluções tradicionais de gerenciamento de redes praticamente não oferecem suporte algum ao gerenciamento baseado na cooperação entre humanos, pois foram concebidas levando em consideração um único domínio administrativo sendo gerenciado por um único operador ou por um grupo de operadores pertencentes a uma mesma organização. Visando solucionar esse problema, é introduzido nesse trabalho o uso de tecnologias peer-to-peer (P2P) em soluções de gerenciamento de redes de modo a permitir o gerenciamento baseado na cooperação entre humanos. Nessa abordagem, operadores de rede usam ferramentas híbridas que, ao mesmo tempo, gerenciam os seus domínios locais e fazem parte de um sistema P2P de gerenciamento. A proposta de gerenciamento apresentada foi implementada através de uma ferramenta chamada ManP2P, a qual combina funcionalidades de gerenciamento de redes e trabalho cooperativo. / Modern networks not rarely require human operators from different administrative domains to interact with one another in order to accomplish a management task in a cooperative fashion. In this scenario one observes that not only management actions are executed in a distributed way, but also the management decisions are cooperatively taken by all autonomous operators participating in a management session. Traditional management solutions fail to support human centered cooperative management because they were conceived considering a single administrative domain being managed by a single operator or by a group of operators belonging to the same organization. In order to address this issue, we introduce in this paper the use of peer-to-peer (P2P) technologies in network management solutions to enable human-centered cooperative management. In this approach, network operators use hybrid tools that both manage their local domains and are peers of an overlay management system. We have implemented the proposed management approach through a tool called ManP2P, which combines network management and groupware functionalities.

O gerenciamento das redes de comunicações modernas frequentemente requer que operadores de diferentes domínios administrativos interajam uns com os outros de modo a realizar uma tarefa de gerenciamento de maneira cooperativa. Nesse contexto, pode-se observar que há distribuição tanto na execução das ações como na tomada das decisões pelos operadores participantes da sessão de gerenciamento. As soluções tradicionais de gerenciamento de redes praticamente não oferecem suporte algum ao gerenciamento baseado na cooperação entre humanos, pois foram concebidas levando em consideração um único domínio administrativo sendo gerenciado por um único operador ou por um grupo de operadores pertencentes a uma mesma organização. Visando solucionar esse problema, é introduzido nesse trabalho o uso de tecnologias peer-to-peer (P2P) em soluções de gerenciamento de redes de modo a permitir o gerenciamento baseado na cooperação entre humanos. Nessa abordagem, operadores de rede usam ferramentas híbridas que, ao mesmo tempo, gerenciam os seus domínios locais e fazem parte de um sistema P2P de gerenciamento. A proposta de gerenciamento apresentada foi implementada através de uma ferramenta chamada ManP2P, a qual combina funcionalidades de gerenciamento de redes e trabalho cooperativo. / Modern networks not rarely require human operators from different administrative domains to interact with one another in order to accomplish a management task in a cooperative fashion. In this scenario one observes that not only management actions are executed in a distributed way, but also the management decisions are cooperatively taken by all autonomous operators participating in a management session. Traditional management solutions fail to support human centered cooperative management because they were conceived considering a single administrative domain being managed by a single operator or by a group of operators belonging to the same organization. In order to address this issue, we introduce in this paper the use of peer-to-peer (P2P) technologies in network management solutions to enable human-centered cooperative management. In this approach, network operators use hybrid tools that both manage their local domains and are peers of an overlay management system. We have implemented the proposed management approach through a tool called ManP2P, which combines network management and groupware functionalities.

This thesis presents the design and evaluation of content-based publish/subscribe systems for efficient content dissemination and sharing of dynamic and large-scale networked applications. The rapid development of network technologies and the continuous investment in network infrastructure have realized a ubiquitous platform for sharing information. However, there lacks efficient protocol and software that can utilize such resource to support novel networked applications. In this thesis, we explore the possibility of content-based publish/subscribe as an efficient communication substrate for dynamic and large-scale networked applications. Although content-based publish/subscribe has been used extensively in many small-to-medium scale systems, there is no Internet-scale applications that utilize this technology. The research reported in this thesis investigates the technical challenges and their solutions of applying content-based publish/subscribe in various applications in mobile networks and Internet. We apply content-based publish/subscribe in the interest-driven information sharing for smartphone networks. We design efficient approximate content matching algorithms and data structures. We study how to construct optimal overlay publish/subscribe overlay networks. We propose architecture designs that make Internet content-based publish/subscribe robust. We also design a name resolution system that enables content discovery in the Internet. These techniques are evaluated comprehensively in realistic simulation studies, and some of them are further evaluated on PlanetLab testbed with prototype implementations. / Computer and Information Science

Les systèmes pair-à-pair (P2P) sont utilisés par des millions d'usagers tous les jours. Dans beaucoup de cas, un usager souhaite communiquer, échanger du contenu ou des services à travers différents systèmes P2P. Cela requiert de la coopération entre les différents systèmes P2P, ce qui est très souvent difficile ou même impossible à obtenir, à cause des raisons suivantes; in primis, l'absence d'une infrastructure de routage entre les réseaux, ce qui rend la communication étanche et, in secundis, l'incompatibilité des protocoles et des opérations des susdits systèmes. L'objectif principal de cette thèse est celui de permettre la coopération entre systèmes P2P. La thèse introduit un cadre de coopération rétro-compatible entre systèmes P2P hétérogènes constitué de deux parties. La première est un cadre de routage intra-réseaux permettant à des réseaux hétérogènes de communiquer. La deuxième est une application coopérative, conçue à l'aide et au travers du cadre de routage, dont l'objectif est celui de résoudre les incompatibilités protocolaires des systèmes P2P sous-jacents. A cause des différences intrinsèques des domaines d'applications - dans la description du cadre de coopération - la thèse se focalise essentiellement sur les principes fondateurs de la construction du cadre de coopération. Comme étude de cas de notre cadre de coopération, on présente une solution complète permettant une coopération entre des réseaux P2P spécialisés dans l'échange des fichiers pouvant s'appliquer aux réseaux P2P actuels. La solution proposée serait composée par une application de coopération, devant fonctionner à l'aide d'un protocole d'échange de fichiers inter-réseaux (IFP), et qui devrait être à la base du cadre de routage inter-protocolaire. A travers des expérimentations en taille réelle, on montre l'efficacité du cadre de routage inter-réseaux ainsi que l'efficacité de la solution proposée, pour permettre la coopération entre réseaux d'échange des fichiers P2P. Pendant l'étude des protocoles de routages inter-réseaux, on a observé que l'interconnexion des systèmes P2P pourrait donner une réponse à des problématiques communes aux systèmes P2P, comme le passage à l'échelle, la localisation, la tolérance aux pannes, et l'auto-organisation. Notre conviction est qu'un système possédant toutes ces caractéristiques pourrait être à la base de la conception de plusieurs systèmes ''cyber-physiques'' (CPS). En fait, dans la deuxième partie de la thèse, on présente une étude de cas d'usage de notre architecture d'interconnexion des réseaux P2P appliquée à un cas typique de CPS, notamment les réseaux d'électricité intelligents. Dans cette étude de cas, une collection des ''tables de hachage distribuées'' et organisées de façon hiérarchique est à la base d'une infrastructure de communication entre de compteurs d'électricité intelligents (AMI). L'analyse et les expérimentations montrent le passage à l'échelle et l'efficacité de communication de l'architecture proposée.

Redes sobrepostas são consideradas uma forma fácil e eficaz de implantar serviços de Internet de valor agregado através da extensão do serviço de entrega de melhor esforço oferecido pela camada de rede IP. No entanto, para aproveitar-se de seus benefícios, um projeto cuidadoso torna-se necessário. Como ainda não existem padrões para redes sobrepostas, com frequência estas são projetadas de forma independente, o que significa que cada uma aplica sua própria descoberta de topologia e toma suas próprias decisões de roteamento. Como resultado, elas podem afetar negativamente umas às outras, através da geração de uma sobrecarga de informações de controle ou através da criação de gargalos na rede física. Portanto, para evitar tais situações, este trabalho propõe uma arquitetura e organização de middleware para serviços em redes sobrepostas. Dentre os possíveis serviços, um esquema de multidifusão sobreposta foi projetado e implantado, no qual um protocolo de middleware baseado em padrões internacionais implementa tanto um cenário intradomínio quanto interdomínio. Validamos o sistema proposto através de dois tipos de análise: uma análise de desempenho e uma análise funcional. A primeira análise mostra que o desempenho do protocolo aqui proposto pode ser comparado ao desempenho dos protocolos tradicionais de multidifusão IP, a menos de uma métrica: a latência. Entretanto, o modelo analítico desenvolvido mostra a possibilidade de se traçar heurísticas para um projeto de serviço sobreposto com latência controlada. Além disso, a análise funcional utiliza-se de medidas experimentais no sistema real para confirmar a funcionalidade de multidifusão interdomínio, facilidade antes inviável em ambientes de rede pública utilizando-se as técnicas tradicionais. / Overlay networks are considered an easy and effective way to deploy value-added Internet services by extending the best-effort delivery service supported by IP layer. However, a careful design is needed in order to take advantage of these benefits. Since there are no standards for overlay networks yet, they are often independently designed, which means that each one applies its own topology discovery and takes its own routing decisions. As a result, they may negatively affect each other by generating control overhead and by creating bottlenecks in the physical network. In order to avoid such a situation, this work proposes an overlay network organization and architecture for overlay network services. Among possible services, an overlay multicast scheme was projected and deployed, on which a middleware protocol based on international standards implements both intradomain and interdomain scenarios. We validate the proposed system by two kind of analysis: performance analysis and functional analysis. The first one shows the performance of the proposed protocol can be compared with performance of traditional IP multicast protocols, except for one metric: latency. Nevertheless, the developed analytical modeling shows the possibility to outline heuristics for a controlled-latency overlay service project. Moreover, functional analysis uses experimental measurements on real system to confirm interdomain multicast functionality, facility before unfeasible in public network environments when using traditional techniques.

Performanz ist ein zentraler Bestandteil des Designs von Anonymisierungsdiensten. Ihre zunehmende Popularität führt jedoch zu einer hohen Netzwerklast, die unzulängliche Entwurfsentscheidungen imminent macht. Die Anforderungen und die vielschichtige Architektur von Anonymisierungsdiensten machen die Thematik zu einem anspruchsvollen und zugleich inspirierenden Forschungsgegenstand. Die vorliegende Arbeit diskutiert das Design von sogenannten Niedriglatenz-Anonymisierungsdiensten im Allgemeinen und dem Tor-Netzwerk als relevantesten Vertreter im Speziellen. Es werden Lösungen für eine Reihe von Forschungsfragen entwickelt, die allesamt das Ziel verfolgen, diese Overlay-Netzwerke zu verbessern und sicherer zu gestalten. Es entsteht ein fundamentales Verständnis zu Netzwerkaspekten in Anonymisierungs-Overlays, das die Netzwerklast, als vorherrschende Ursache für die schwache Performanz, thematisiert. / Performance is a pivot point in the design of anonymity overlays. Due to their growing popularity, they are faced with increasing load, which makes design problems imminent. The special requirements and complex architecture of anonymity overlays renders the topic a challenging but likewise inspiring object of research. In this work, we discuss the design of low-latency anonymous communication systems in general and the Tor network as the de-facto standard in particular. We develop solutions to a number of research questions, all collectively following the aim of enhancing and securing such networks. By doing this we create a fundamental technical understanding of networking aspects in anonymity overlays and tackle the most prevalent performance issue experienced today: network congestion.

Le routage par clé et par contenu sont des systèmes de routage ou la destination d'un message suit un parcours entre les nœuds du réseau qui dépend seulement du contenu du message même. On peut les trouver utilisés soit dans des systèmes pair-à-pair connus comme Réseaux Overlay Structurés (Structured Overlay Networks, SON), soit dans les architecture internet de nouvelle génération, les Réseaux Centrés sur les Contenus (Content-Centric Networks, CCN). Le but de cette thèse est double. D'un côté, on explore le sujet de l'interconnexion et de la coopération des réseaux d'overlay, et on propose une architecture capable de permettre à plusieurs réseaux d'overlay hétérogènes, avec différentes topologies et différents mécanismes de routage, d'interagir, grâce à une infrastructure basée sur des nœuds passerelles. On montre, par des moyennes de simulation et déploiement dans un réseaux réel, que la solution est scalable et permet un routage quasi-exhaustif avec un nombre relativement bas des nœuds passerelle bien connectés. De plus, on présente deux exemples d'applications qui pourront bénéficier de cette architecture. Dans une deuxième partie, on rentre plutôt dans les possibilités offertes par le routage basé sur les contenus hors sa "zone de confort": d'abord, on analyse les améliorations qu'un réseau d'overlay structuré peut porter à un système de diffusion vidéo pair-à-pair, en termes de qualité du vidéo et de perte des paquets pendant la transmission. Après, on examine un système entièrement centré sur le routage basé sur les contenus, en développant une solution de diffusion vidéo en temps réel dans un réseau CCN. / Key-based and content-based routing are a class of routing techniques where the destination and routing path for a message depends solely on the content of the message itself. This kind of routing has been implemented in certain peer-to-peer systems, known as Structured Overlay Networks (SON), or in the Next Generation Internet architectures, under the name of Content-Centric Networks (CCN). The scope of this thesis is twofold: on the one side, we explore the topic of the interconnection and cooperation of different structured overlays, and propose architecture capable of allowing several heterogeneous overlay networks, with different topologies and different routing schemes, to interact, thanks to a lightweight infrastructure consisting of co-located nodes. Through the use of simulations and real-world deployment, we show how this solution is scalable and how it facilitates quasi-exhaustive routing, with even a relatively low number of well-connected co-located nodes. To address the problem of scaling network design to millions of nodes, we propose a mathematical model capable of deriving basic performance figures for an interconnected system. Furthermore, we present two application examples that could greatly benefit from such architecture. On the other side, we investigate a little further into the capabilities of content-based routing outside of its "comfort zone": first, we analyze the improvement that a SON could bring to a peer-to-peer real-time video streaming system (P2P-TV), in terms of chunk loss and Quality of Experience. Then, we move the approach to a fully content-based domain, implementing the P2P-TV solution on top of Content-Centric Networks.

Redes sobrepostas são consideradas uma forma fácil e eficaz de implantar serviços de Internet de valor agregado através da extensão do serviço de entrega de melhor esforço oferecido pela camada de rede IP. No entanto, para aproveitar-se de seus benefícios, um projeto cuidadoso torna-se necessário. Como ainda não existem padrões para redes sobrepostas, com frequência estas são projetadas de forma independente, o que significa que cada uma aplica sua própria descoberta de topologia e toma suas próprias decisões de roteamento. Como resultado, elas podem afetar negativamente umas às outras, através da geração de uma sobrecarga de informações de controle ou através da criação de gargalos na rede física. Portanto, para evitar tais situações, este trabalho propõe uma arquitetura e organização de middleware para serviços em redes sobrepostas. Dentre os possíveis serviços, um esquema de multidifusão sobreposta foi projetado e implantado, no qual um protocolo de middleware baseado em padrões internacionais implementa tanto um cenário intradomínio quanto interdomínio. Validamos o sistema proposto através de dois tipos de análise: uma análise de desempenho e uma análise funcional. A primeira análise mostra que o desempenho do protocolo aqui proposto pode ser comparado ao desempenho dos protocolos tradicionais de multidifusão IP, a menos de uma métrica: a latência. Entretanto, o modelo analítico desenvolvido mostra a possibilidade de se traçar heurísticas para um projeto de serviço sobreposto com latência controlada. Além disso, a análise funcional utiliza-se de medidas experimentais no sistema real para confirmar a funcionalidade de multidifusão interdomínio, facilidade antes inviável em ambientes de rede pública utilizando-se as técnicas tradicionais. / Overlay networks are considered an easy and effective way to deploy value-added Internet services by extending the best-effort delivery service supported by IP layer. However, a careful design is needed in order to take advantage of these benefits. Since there are no standards for overlay networks yet, they are often independently designed, which means that each one applies its own topology discovery and takes its own routing decisions. As a result, they may negatively affect each other by generating control overhead and by creating bottlenecks in the physical network. In order to avoid such a situation, this work proposes an overlay network organization and architecture for overlay network services. Among possible services, an overlay multicast scheme was projected and deployed, on which a middleware protocol based on international standards implements both intradomain and interdomain scenarios. We validate the proposed system by two kind of analysis: performance analysis and functional analysis. The first one shows the performance of the proposed protocol can be compared with performance of traditional IP multicast protocols, except for one metric: latency. Nevertheless, the developed analytical modeling shows the possibility to outline heuristics for a controlled-latency overlay service project. Moreover, functional analysis uses experimental measurements on real system to confirm interdomain multicast functionality, facility before unfeasible in public network environments when using traditional techniques.