La dépendance des différentes infrastructures vis-à-vis du réseau Internet va croissant. D’abord la convergence des médias mais bientôt l’Internet des objets ou les véhicules autonomes peut-être vont contribuer à augmenter la criticité d’Internet. Il est donc important de résoudre les problèmes liés à l’infrastructure actuelle, en terme de passage à l’échelle, de confidentialité ou bien de fiabilité. Les communications multichemins font partie des possibilités pour attaquer ce défi. Pour autant la transition vers ces technologies n’est pas sans difficulté. En effet certains équipements bloquent les protocoles inconnus, empêchant ainsi l’émergence de nouvelles technologies. C’est un phénomène en partie responsable de l’ossification d’internet. D’autres considérations techniques limitent l’intérêt de recourir à des technologies multichemins dans certains cas, puisque celles-ci peuvent alors présenter des performances moindres que les technologies monochemins. Dans le cadre de cette thèse, nous proposons des réponses à certains de ces cas afin de maximiser le spectre d’application des technologies multichemins, en particulier du protocole Multipath TCP (MPTCP). Plus précisémement, après une revue détaillée du domaine des communications multichemins, nous proposons une réponse au problème de découverte des chemins. De plus, motivés par l’intuition que les ordonnanceurs peuvent s’appuyer sur les latences unidirectionelles, afin de lutter contre l’arrivé de paquets dans le désordre, nous proposons une technique qui ne modifie que l’envoyeur de données pour estimer cette métrique. En outre, nous proposons un outil qui maximise le débit tout en prenant en compte des politiques utilisateur par exemple pour forcer l’envoi d’une partie du trafic sur un chemin peu performant mais qui va coûter moins cher à l’utilisateur. Finalement, nous développons et évaluons un modèle de MPTCP. / The criticity of the Internet keeps increasing with a very high number of services depending on its infrastructure. The Internet is expected to support services with an increasing tangible impact on the physical world such as the Internet of Things (IoT) or autonomous vehicles. It is thus important to address the current infrastructure shortcomings in terms of scalability, confidentiality and reliability. Multipath communications are one possible solution to address this challenge. The transition towards multipath technologies is not obvious, there are several challenges ahead. Some network devices block unknown protocols, thus preventing the emergence of new technologies, which plays a part in what is often referred to as the ossification of the Internet. Moreover, due to technical reasons, there are cases for which multipath technologies perform worse than their single path counterpart. In this thesis, we are interested in addressing some of these cases and limit their impact, so that multipath communications perform better than single path communications as often as possible. More specifically, we propose enhancements to Multipath TCP (MPTCP). After a detailed survey of multipath communications across all layers, we propose an answer as to the question of how many paths to use and how to ensure proper forwarding. Moreover, motivated by the intuition that packet arrival disorder can be mitigated by the knowledge of one way latencies, we propose a latency estimator with sender-side modifications only. Furthermore, as throughput maximization is in general solved regardless of the interface cost or user preferences, we elaborate a framework capable of presenting more complex strategies if for instance the user wants to enforce throughput even on less efficient paths. Finally, we develop and present a complete simulation model of MPTCP.

Modern mobile phones often have multiple network interfaces. By utilising these interfaces concurrently, the user experience can be enhanced, e.g., increased performance, stability and reliability. The introduction of 5G has increased the theoretical bandwidth and lowered the latency compared to previous generations of cellular techniques. However, this leads to a larger parameter space, which in a multipath system, increases the likelihood of asymmetric paths. Path asymmetry affects the overall performance of a reliable multipath connection, due to, e.g.,variation in delay which can cause packets to arrive out-of-order. This thesis first explores how multipath systems perform in comparison to single-path systems. Secondly, when the simultaneous use of paths is not beneficial for 5G-like multipath systems, and finally, if its possible to translate the findings into an effective scheduling strategy to improve the performance. With the use of Mininet, a network emulator, path parameters relating to 3G, 4G, 5G and WiFi are used to generate results representing a large spectrum of path combinations. The results gives insight to how MPTCP performs, from symmetric to highly asymmetric paths. The results show that the usage of multiple paths is not beneficial when the round-trip times are of different orders-of-magnitude. Based on observations, a proof-of-concept scheduler is proposed that reduce the negative effect of path asymmetry while still performing low amount of calculations.

Même si les réseaux de données ont beaucoup évolué au cours des dernières décennies, les paquets sont presque toujours transmis d’un nœud à l’autre comme des blocs de données inaltérables. Cependant, ce paradigme fondamental est aujourd’hui remis en question par des techniques novatrices comme le codage réseau, qui promet des améliorations de performance et de fiabilité si les nœuds sont autorisés à mixer des paquets entre eux. Les réseaux sans fil manquent de fiabilité en raison des obstacles ou interférences que subissent les liens sans fil, et ces problèmes peuvent empirer dans des topologies maillées avec de multiples relais potentiels. Dans ce travail, nous nous concentrons sur l’application du codage réseau intra-flux aux flux unicast dans les réseaux sans fil, avec pour objectif d’améliorer la fiabilité des transferts de données et de discuter des opportunités de déploiement et des performances. Tout d’abord, nous proposons une borne inférieure pour la redondance, puis un algorithme opportuniste distribué, pour adapter le codage aux conditions du réseau et permettre la livraison fiable des données dans un réseau sans fil maillé, tout en prenant en compte les besoins de l’application. En outre, puisque les opérations requises pour le codage réseau sont coûteuses en termes de calcul et de mémoire, nous étendons cet algorithme pour s’adapter aux contraintes physiques de chaque nœud. Ensuite, nous étudions les interactions du codage intra-flux avec TCP et son extension MPTCP. Le codage réseau peut en effet améliorer les performances de TCP, qui ont tendance à être plus faibles sur les liens sans fil, moins fiables. Nous observons l’impact des problèmes d’équité qui se posent quand des flux codés fonctionnent en parallèle avec des flux traditionnels non codés. Pour finir, nous explorons deux manières différentes d’améliorer les performances de MPTCP dans les environnements sans fil : le faire fonctionner sur du codage réseau, et implémenter directement le codage directement dans le protocole MPTCP tout en préservant sa compatibilité avec TCP / Even if packet networks have significantly evolved in the last decades, packets are still transmitted from one hop to the next as unalterable pieces of data. Yet this fundamental paradigm has recently been challenged by new techniques like network coding, which promises network performance and reliability enhancements provided nodes can mix packets together. Wireless networks rely on various network technologies such as WiFi and LTE. They can however be unreliable due to obstacles, interferences, and these issues are worsened in wireless mesh network topologies with potential network relays. In this work, we focus on the application of intra-flow network coding to unicast flows in wireless networks. The main objective is to enhance reliability of data transfers over wireless links, and discuss deployment opportunities and performance. First, we propose a redundancy lower bound and a distributed opportunistic algorithm, to adapt coding to network conditions and allow reliable data delivery in a wireless mesh. We believe that application requirements have also to be taken into account. Since network coding operations introduce a non negligible cost in terms of processing and memory resources, we extend the algorithm to consider the physical constraints of each node. Then, we study the interactions of intra-flow coding with TCP and its extension MPTCP. Network coding can indeed enhance the performances of TCP, which tends to perform poorly over lossy wireless links. We investigate the pratical impact of fairness issues created when running coded TCP flows besides legacy non-coded TCP flows. Finally, we explore two different ways to enhance the performance of MPCTP in wireless environments : running it over network coding, and implementing the coding process directly in MPTCP while keeping it fully TCP-compatible.

Wireless communication has seen a tremendous growth in the last decades. Continuing on this trend, wireless multi-hop networks  are nowadays used or planned for use in a multitude of contexts, spanning from Internet access at home to emergency situations. The Transmission Control Protocol (TCP) provides reliable and ordered delivery of a data and is used by major Internet applications such as web browsers, email clients and file transfer programs. TCP traffic is also the dominating traffic type on the Internet. However, TCP performs less than optimal in wireless multi-hop networks due to packet reordering, low link capacity, packet loss and variable delay. In this thesis, we develop novel proposals for enhancing the network and transport layer to improve TCP performance in wireless multi-hop networks. As initial studies, we experimentally evaluate the performance of different TCP variants, with and without mobile nodes. We further evaluate the impact of multi-path routing on TCP performance and propose packet aggregation combined with aggregation aware multi-path forwarding as a means to better utilize the available bandwidth. The last contribution is a novel extension to multi-path TCP to  enable single-homed hosts to fully utilize the network capacity. / <p>Opponent changed. Prof. C. Lindeman from the University of Leipzig was substituted by Prof. Zhang.</p>

Submitted by Milena Rubi (milenarubi@ufscar.br) on 2017-06-01T14:58:47Z No. of bitstreams: 1 SILVA_Alan_2016.pdf: 14722594 bytes, checksum: 0c8346924c434318f6c349f7ed8112d9 (MD5) / Approved for entry into archive by Milena Rubi (milenarubi@ufscar.br) on 2017-06-01T14:58:54Z (GMT) No. of bitstreams: 1 SILVA_Alan_2016.pdf: 14722594 bytes, checksum: 0c8346924c434318f6c349f7ed8112d9 (MD5) / Approved for entry into archive by Milena Rubi (milenarubi@ufscar.br) on 2017-06-01T14:58:59Z (GMT) No. of bitstreams: 1 SILVA_Alan_2016.pdf: 14722594 bytes, checksum: 0c8346924c434318f6c349f7ed8112d9 (MD5) / Made available in DSpace on 2017-06-01T14:59:06Z (GMT). No. of bitstreams: 1 SILVA_Alan_2016.pdf: 14722594 bytes, checksum: 0c8346924c434318f6c349f7ed8112d9 (MD5) Previous issue date: 2016-12-06 / Não recebi financiamento / The Multipath TCP (MPTCP) protocol allows applications to better explore the network resources available to multi-connected devices such as mobile phones or multi-homed systems. Here, some advantages are envisioned: bandwidth aggregation, the ability to maintain the connection, if one of the network path fails and the use of multiple paths. To extend these capabilities to the application, RFC 6897 defines an API to better control each of MPTCP’s subflows, so that these can be added or removed as needed. This work presents an initial API implementation as defined in RFC 6897. We implemented some functions described in the document, such as protocol on/o, check existent subflows and add new subflows. To test the API and validate our implementation we built an HTTP application that detects elephant flows and uses the API for open new subflows using the original TCP connection. Some tests were performed in a network using a cubic topology and showed that the API utilization decreased the Flow Completion time of TCP connections. / O protocolo Multipath TCP (MPTCP) permite que as aplicações possam explorar melhor os recursos de rede disponíveis para dispositivos multiconectados como os telefones móveis ou sistemas multi-homed. Aqui, algumas vantagens são previstas: agregação de banda, a habilidade de manter a conexão estabelecida se houver falha em um dos caminhos de rede e a utilização de múltiplos caminhos. Para estender essas capacidades para a aplicação, a RFC 6897 define uma API que permite um melhor controle de cada subfluxo MPTCP, de modo que esses possam ser adicionados ou removidos conforme necessário. Este trabalho apresenta uma implementação inicial da API descrita na RFC 6897 para o protocolo MPTCP. Sendo assim, implementamos algumas das funções de manipulação do protocolo MPTCP descritas no documento, quais sejam: ligar e desligar o protocolo, verificar subfluxos existentes e adicionar novos subfluxos. Para testar a API e validar a nossa implementação, nós desenvolvemos uma aplicação HTTP que detecta fluxos elefantes e utiliza a API para abrir novos subfluxos a partir da conexão TCP original. Testes de desempenho foram realizados em uma topologia cúbica e mostraram que a utilização da API pela aplicação diminuiu o Flow Completion Time das conexões TCP.

Innovations in optical communication are expected to transform the landscape of global communications, internet and datacenter networks. This dissertation investigates several important issues in optical communication such as fairness, throughput, blocking probability and differentiated quality of service (QoS). Novel algorithms and new approaches have been presented to improve the performance of optical circuit switching (OCS) and optical burst switching (OBS) for long haul, and datacenter networks. Extensive simulations tests have been conducted to evaluate the effectiveness of the proposed algorithms. These simulation tests were performed over a number of network topologies such as ring, mesh and U.S. Long-Haul, some high processing computing (HPC) topologies such as 2D and 6D mesh torus topologies and modern datacenter topologies such as FatTree and BCube. Two new schemes are proposed for long haul networks to improve throughput and hop count fairness in OBS networks. The idea is motivated by the observation that providing a slightly more priority to longer bursts over short bursts can significantly improve the throughput of the OBS networks without adversely affecting hop-count fairness. The results of extensive performance tests have shown that proposed schemes improve the throughput of optical OBS networks and enhance the hop-count fairness. Another contribution of this dissertation is the research work on developing routing and wavelength assignment schemes in multimode fiber networks. Two additional schemes for long haul networks are presented and evaluated over multimode fiber networks. First for alleviating the fairness problem in OBS networks using wavelength-division multiplexing as well as mode-division multiplexing while the second scheme for achieving higher throughput without sacrificing hop count fairness. We have also shown the significant benefits of using both mode division multiplexing and wavelength division multiplexing in real-life short-distance optical networks such as the optical circuit switching networks used in the hybrid electronic-optical switching architectures for datacenters. We evaluated four mode and wavelength assignment heuristics and compared their throughput performance. We also included preliminary results of impact of the cascaded mode conversion constraint on network throughput. Datacenter and high performance computing networks share a number of common performance goals. Another highly efficient adaptive mode wavelength- routing algorithm is presented over OBS networks to improve throughput of these networks. The effectiveness of the proposed model has been validated by extensive simulation results. In order to optimize bandwidth and maximize throughput of datacenters, an extension of TCP called multipath-TCP (MPTCP) has been evaluated over an OBS network using dense interconnect datacenter topologies. We have proposed a service differentiation scheme using MPTCP over OBS for datacenter traffic. The scheme is evaluated over mixed workload traffic model of datacenters and is shown to provide tangible service differentiation between flows of different priority levels. An adaptive QoS differentiation architecture is proposed for software defined optical datacenter networks using MPTCP over OBS. This scheme prioritizes flows based on current network state.