Cette thèse s´intéresse à la conception et l´évaluation des protocoles de routage et d´accès au canal pour les réseaux sans fils. La première partie de la thèse focalise principalement sur l´évaluation de protocoles de routage pour les réseaux tolérants aux perturbations quand ces réseaux incluent des relais fixes, appelées boîtes. Dans un premier temps, nous montrons que les instants successifs de rencontre entre une boîte et un noeud mobile qui se déplace selon un modèle de mobilité aléatoire sont bien approximés par un processus de Poisson. Nous donnons une formule explicite approchée pour l´intensité de ce processus qui dépend notamment de la densité de probabilité spatiale du modèle de mobilité considérée ainsi que celle des boîtes. Dans un deuxième temps, nous étudions l´impact d´ajouter des boîtes sur les performances de deux protocoles de routage classiques, le protocole épidémique et le protocole de routage à deux sauts. Nous développons des expressions explicites pour quantifier la distribution et la moyenne du délai de livraison d´un paquet, ainsi que le nombre des copies générées lors de cette transmission. Ensuite, nous proposons cinq stratégies qui s´appuient sur la présence des boîtes pour réaliser le routage des copies. Par ailleurs, nous introduisons une plateforme basée sur un modèle markovien qui permet de calculer et de comparer analytiquement les diverses métriques de performance pour ces cinq stratégies. Dans la deuxième partie de la thèse, nous intéressons à l´algorithme de backoff du standard IEEE 802.11. Nous proposons une extension de cet algorithme dont l´objectif est d´améliorer ses performances dans le cas où le réseau possède un grand nombre d´utilisateurs.

La presencia de topologías multi-salto en comunicaciones inalámbricas de todo tipo es cada vez más apreciable, esperándose además que esta tendencia se mantenga en un futuro cercano. A pesar de que inicialmente fueran concebidos para solventar la falta de infraestructura subyacente en ciertos escenarios concretos, estos despliegues han acaparado el interés de diferentes actores del ámbito de las comunicaciones (incluyendo los operadores), con lo que es razonable pensar que su relevancia irá creciendo paulatinamente. De hecho, existen diversas iniciativas en algunos foros de estandarización que de alguna manera corroboran este hecho. Hay que tener en cuenta, por otro lado, otros factores adicionales (como la eclosión que se ha producido en el campo de las redes de sensores inalámbricos) que seguramente fomentarán el uso de estas topologías.A pesar de la creciente actividad en el ámbito de los despliegues multi-salto, sigue siendo necesario establecer, de manera cuantitativa, cuáles son sus posibles beneficios, tanto para los usuarios finales de los sistemas de comunicación, como para los operadores, especialmente teniendo en cuenta el elevado grado de heterogeneidad que también caracterizará las redes inalámbricas.Por otro lado, en lo que se refiere a los algoritmos y protocolos a ser empleados sobre este tipo de topologías, y a pesar de la intensa labor de investigación que sobre ellos se ha realizado recientemente, queda aún un número relevante de aspectos a analizar. En primer lugar, el mero hecho de que sus requerimientos y retos iniciales hayan variado de manera sustancial puede, y debe, influenciar sus principios básicos. Además, es necesario acometer verificaciones de los mismos sobre plataformas reales y, así mismo, asegurar que las evaluaciones que se realicen con técnicas de simulación utilicen modelos reales que permitan reflejar de manera fidedigna las condiciones que se dan en la realidad.Esta Tesis afronta, en primer lugar, la evaluación cuantitativa de la mejora que es posible alcanzar al utilizar topologías inalámbricas multi-salto para extender despliegues de red más tradicionales. Un primer aspecto que es razonable considerar es la ampliación de la cobertura que se consigue; en este caso se ha realizado un análisis que sigue un doble enfoque, analítico y mediante técnicas de simulación, para determinar es la ganancia que se logra. Se ha partido de dos modelos de red complementarios entre sí, asumiendo en el primero de ellos una falta total de planificación previa, mientras que en el segundo se utiliza un emplazamiento óptimo de los elementos de conexión a la red. Se ha comprobado que, a pesar de sus características claramente antagónicas, los resultados obtenidos con ambos escenarios son similares entre sí. Además, se concluye que, a pesar de que la ampliación de la cobertura que se alcanza es muy relevante, es posible establecer un límite razonable para el número máximo de saltos a emplear, ya que la mejoría adicional al incrementar la longitud de la ruta deja de ser apreciable a partir del mismo. Este aspecto se podría aprovechar para influir en el diseño de las técnicas de encaminamiento a emplear sobre este tipo de topologías.Se analizan también otras mejoras adicionales, utilizando un escenario en el que predomina sobremanera la heterogeneidad de los elementos de conexión a la red, e integrando las extensiones multi-salto con un algoritmo de selección de acceso genérico, que permite modular el peso que tienen las diferentes entidades (tanto el terminal de usuario como la propia red), así como diversos parámetros y restricciones a considerar, a la hora de determinar la alternativa de conexión óptima. Se comprueba que, tanto para los usuarios, gracias a la mejora de la calidad del servicio que perciben, como para la red, que incrementa el tráfico que es capaz de cursar, el uso de comunicaciones multi-salto puede resultar altamente atractivo.En segundo lugar, también se afronta la mejora de las técnicas de encaminamiento que tradicionalmente se emplean sobre las redes multi-salto, basadas en minimizar el número de saltos entre los dos extremos de la comunicación. Para ello, y utilizando el paradigma de Cross-Layer Optimisation, se propone una versión mejorada del protocolo DSR, denominada SADSR, que emplea la información relativa a la calidad de los enlaces subyacentes para modular el algoritmo de selección de ruta. Destacar que se acomete una verificación experimental de dicha propuesta, contestando, de este modo, a una de las reivindicaciones más importantes en este campo. Los resultados obtenidos permiten inferir que las prestaciones de la propuesta realizada en el marco de esta Tesis son superiores a los de la versión original del DSR.A pesar del evidente valor que las validaciones empíricas aportan, tienen la limitación de que es complicado establecer topologías con un número elevado de nodos, o acometer experimentos con una pauta repetitiva para determinar un comportamiento promedio. Es por ello que también se acomete un análisis mediante técnicas de simulación, en el que se compara SADSR con la versión original del DSR, así como con otras propuestas que han acaparado el interés por parte de la comunidad científica. Los resultados que se derivan a raíz de la evaluación llevada a cabo no hacen sino corroborar lo que ya adelanta la verificación experimental, ya que las prestaciones del SADSR son muy superiores a las del resto de alternativas analizadas.Para llevar a cabo el análisis anterior es fundamental el uso de un modelo de canal que refleje, con el mayor grado de exactitud posible, un comportamiento realista. Para ello, en la Tesis se afronta el diseño, implementación y posterior integración en la plataforma Network Simulator de BEAR, un modelo de canal basado en filtrado auto-regresivo, que se caracteriza principalmente por emular la aparición de errores a ráfagas que se observa en entornos reales. En ese sentido, se parte de un extenso conjunto de medidas que permite corroborar el correcto funcionamiento de la propuesta. / The presence of multi-hop topologies within all types of wireless communications is becoming more and more common, and this tendency is expected to be maintained in the near future. Although they were originally conceived to compensate the lack of subjacent infrastructure in certain scenarios, these deployments have attracted the interest of different actors in the wireless communications value chain (including network operators) and thus it is logical to think that their relevance will gradually increase. In fact, there already exist some standardization initiatives which corroborate this point to some extent. Furthermore, other additional factors, such as the rapid growth which has been seen in wireless sensor technologies, also strengthen the use of these topologies..In spite of the growing activity in the multi-hop deployment field, it is still necessary to establish, in a quantitative way, their potential benefits, both for the end-users of the communication systems, as well as for the operators, considering, in addition, the high degree of heterogeneity which will characterize wireless networks in the future.On the other hand, as far as algorithms and protocols to be used over this type of topology are concerned, and despite the intense research which has been conducted into them, there is still a large number of issues to be tackled. First, the simple fact that their initial requirements and challenges have been modified can, and must, influence their basic principles. In addition, it becomes necessary to address their validation on real platforms and, on the other hand, to ensure that simulation-based evaluations of their performance make use of realistic models which accurately reflect the conditions which are observed in real scenarios.This dissertation tackles, on the one hand, the quantitative evaluation of the improvements which are achievable when using multi-hop topologies to extend legacy network deployments. One first aspect which is logical to consider is the increase in the coverage which is brought about. In this sense, a two-fold approach has been followed, employing both an analytical as well as a simulation-based analysis, to establish what the gain is. Two network models have been used, being complementary to each other; the first one assumes a complete lack of network planning for the deployment of the access elements, while the second one assumes an optimum distribution of them. Although their characteristics are completely different, the results are somehow similar for both cases. Furthermore, it can be concluded that, despite the coverage extension which can be obtained, it is indeed possible to establish a reasonable limit on the maximum number of hops to be used, since the improvement becomes less relevant for higher values. This aspect could influence the design of routing techniques to be used over this type of topology.Furthermore, other additional benefits have been also analyzed, using a network deployment in which the presence of heterogeneity (multi-access) is evident. The multi-hop extensions have been integrated within a generic access selection algorithm which enables the modification of the weights which are assigned to the different entities (both the end-user terminals and the network) as well as to the set of parameters and constraints to be considered when selecting the most appropriate access alternative. It is concluded that for both the end-users, who improve their perception of the quality of service, and the network, which is able to increase the overall amount of traffic possible to be handled, multi-hop extensions are certainly beneficial.In addition, the dissertation also tackles the improvement of the routing techniques which are traditionally employed over multi-hop networks, which are based on minimizing the number of hops between the two sides of the communication. To accomplish this, and using the Cross-Layer Optimisation paradigm, an improved version of the DSR protocol is proposed, namely SADSR. It uses information about the subjacent link qualities to modulate the route selection algorithm. It is worth highlighting that a fully empirical (on a real platform) validation has been conducted, addressing one of the most pressing demands within this field. The results obtained allow us to infer that the proposal made in the framework of this dissertation is clearly outperforming the original DSR version.Although the added value provided by empirical validations is unquestionable, they also have some limitations. First, they normally do not favour the establishment of large topologies, or to undertake repetitive experiments aimed at finding an average behaviour. In this sense, a simulation-based analysis is also used so as to compare the SADSR with the original DSR version as well as with other proposals which have recently attracted interest from the scientific community. The results obtained by the validation carried out on a real platform are confirmed, since the performance brought about by the SADSR is somewhat higher than that of the other strategies.In order to perform the previous analysis the use of a realistic channel model, able to capture with a high degree of accuracy the behaviour exhibited on real platforms, is mandatory. To fulfil this requirement, the dissertation also tackles the design, implementation, and integration within the Network Simulator platform of BEAR, a channel model based on auto-regressive filtering. It is mainly characterized by being able to emulate the bursty presence of errors which is observed over real channels. The design is based on an extensive set of measurements which is used to assess the validity of the proposal.

La thèse se focalise sur la sécurité dans les réseaux mobiles ad hoc (MANET : Mobile Ad hoc NETwork) [RFC 2501]. L'absence d'une gestion centrale des fonctionnalités du réseau rend ces réseaux beaucoup plus vulnérables aux attaques que les réseaux sans fil (WLAN) et filaires (LAN). Malheureusement, les protocoles de sécurité qui existent actuellement ne sont pas conçus pour un tel environnement (dynamique). Ils ne prennent pas la contrainte des ressources en considération car non seulement l'environnement est dynamique, mais les ressources sont aussi limitées (mémoire, capacité de calcul et surtout énergie), ce qui complique davantage la problématique, car on sait bien que les solutions de sécurité sont gourmandes en terme de ressources. Cependant, en raison de l'importance des domaines d'application des réseaux mobiles ad hoc comme les opérations militaires (communication entre les avions, les voitures et le personnel et opérations de secours, situations d'urgence en cas de sinistre, etc . . .), il faut relever le défi, car concevoir un mécanisme de sécurité infaillible pour les réseaux mobiles ad hoc est nécessaire. L'objectif principal de la thèse consiste à étudier les solutions susceptibles d'assurer la sécurité dans les réseaux mobiles ad hoc, en proposant une architecture hiérarchique distribuée qui permet d'établir une infrastructure dynamique à clé publique. Cette architecture doit supporter les différentes caractéristiques de ces réseaux (absence d'une unité centrale de gestion de réseau, topologie réseau dynamique, etc . . .). Dans ce but, un modèle de confiance adapté à l'environnement dynamique pour assurer l'évolution des niveaux de confiance des nœuds est établi. De plus, les vulnérabilités au niveau des autorités de certification sont prises en compte dans le nouveau concept de DDMZ (zone dynamique démilitarisée) que nous proposons. Dans le but de sécuriser les nœuds dont le rôle est crucial au sein du réseau, leur identité doit être cachée. C'est pourquoi le concept d'anonymat est introduit. Un protocole d'authentification anonyme est proposé. De plus, nous nous inspirons du modèle militaire pour mettre en place un mécanisme de camouflage qui cache le rôle des nœuds sensibles. Pour entretenir le modèle de confiance, un mécanisme de surveillance est indispensable. Il est adapté aux contraintes de l'environnement sans fil dynamique et réduit le taux de fausses alarmes (faux positifs). Il est fondé sur une approche inter-couches et un modèle probabiliste pour améliorer l'observation du nœud surveillant. Pour faire face aux attaques intelligentes de type inter-couches, une étude des vulnérabilités au niveau des couches inférieures comme la couche MAC est menée. Ensuite, des mécanismes de prévention et de détection sont analysés et évalués. La performance de ces mécanismes est évaluée avec la prise en compte des métriques primordiales pour les réseaux mobiles ad hoc, telles que la consommation d'énergie, la mobilité, la densité des nœuds et du trafic, etc . . .

On considère souvent IEEE 802.11 comme la technologie sous-jacente aux réseaux sans fil multisauts. Pourtant, dans de tels réseaux, 802.11 s'avère inefficace et/ou inéquitable dans de nombreuses situations. Cette thèse décrit un algorithme d'allocation dynamique et distribuée de débit permettant de garantir un partage équitable de la bande passante entre flux, tout en assurant une utilisation efficace du réseau. De plus, cette thèse propose une nouvelle méthode d'évaluation de l'efficacité d'un protocole en prenant en compte l'expérience utilisateur.

This dissertation applies theories and concepts of wireless communications and signal processing to the security domain to assess the security of a Wireless secret Key Generation (WKG) system against capable eavesdroppers, who employ all the feasible tools to compromise the system’s security. The security of WKG is evaluated via real wireless measurements, where adversary knows and applies appropriate signal processing tools in ordere to predict the generated key with the communicating pair. It is shown that in a broadband stationary wireless communication channel, (e.g. commercial off-the-shelf 802.11 WLAN devices), a capable eavesdropper can recover a large portion of the secret key bits. However, in an Ultra-wideband (UWB) communication, at the same stationary environment, secret key rates of 128 bits per channel probe are achievable. / Graduate

This research investigates wireless intrusion detection techniques for detecting attacks on IEEE 802.11i Robust Secure Networks (RSNs). Despite using a variety of comprehensive preventative security measures, the RSNs remain vulnerable to a number of attacks. Failure of preventative measures to address all RSN vulnerabilities dictates the need for a comprehensive monitoring capability to detect all attacks on RSNs and also to proactively address potential security vulnerabilities by detecting security policy violations in the WLAN. This research proposes novel wireless intrusion detection techniques to address these monitoring requirements and also studies correlation of the generated alarms across wireless intrusion detection system (WIDS) sensors and the detection techniques themselves for greater reliability and robustness. The specific outcomes of this research are: A comprehensive review of the outstanding vulnerabilities and attacks in IEEE 802.11i RSNs. A comprehensive review of the wireless intrusion detection techniques currently available for detecting attacks on RSNs. Identification of the drawbacks and limitations of the currently available wireless intrusion detection techniques in detecting attacks on RSNs. Development of three novel wireless intrusion detection techniques for detecting RSN attacks and security policy violations in RSNs. Development of algorithms for each novel intrusion detection technique to correlate alarms across distributed sensors of a WIDS. Development of an algorithm for automatic attack scenario detection using cross detection technique correlation. Development of an algorithm to automatically assign priority to the detected attack scenario using cross detection technique correlation.

In recent years, there has been a rapid growth in deployment and usage of realtime network applications, such as Voice-over-IP, video calls/video conferencing, live network seminars, and networked gaming. The continued increase in the popularity of realtime applications requires a more intense focus on the provision of strict guarantees for Quality of Service (QoS) parameters such as delay, jitter and packet loss in access networks. At the same time, wireless networking technologies have become increasingly popular with a wide array of devices such as laptop computers, Personal Digital Assistants (PDAs), and cellular phones being sold with built-in WiFi and WiMAX interfaces. For realtime applications to be popular over wireless networks, simple, robust and effective QoS mechanisms suited for a variety of heterogeneous wireless networks must be devised. Implementing the same QoS mechanisms across multiple neighbouring networks aids seamless handover by ensuring that a flow will be treated in the same way, both before and after handover. To provide guaranteed QoS, an access network should limit load using an admission control algorithm. In this research, we propose a method to provide effective admission control for variable bit rate realtime flows, based on the Central Limit Theorem. Our objective is to estimate the percentage of packets that will be delayed beyond a predefined delay threshold, based on the mean and variance of all the flows in the system. Any flow that will increase the percentage of delayed packets beyond an acceptable threshold can then be rejected. Using simulations we have shown that the proposed method provides a very effective control of the total system load, guaranteeing the QoS for a set of accepted flows with negligible reductions in the system throughput. To ensure that flow data is transmitted according to the QoS requirements of a flow, a scheduling algorithm must handle data intelligently. We propose methods to allow more efficient scheduling by utilising existing Medium Access Control mechanisms to exchange flow information. We also propose a method to determine the delay-dependent "value" of a packet based on the QoS requirements of the flow. Using this value in scheduling is shown to increase the number of packets sent before a predetermined deadline. We propose a measure of fairness in scheduling that is calculated according to how well each flow's QoS requirements are met. We then introduce a novel scheduling paradigm, Delay Loss Controlled-Earliest Deadline First (DLC-EDF), which is shown to provide better QoS for all flows compared to other scheduling mechanisms studied. We then study the performance of our admission control and scheduling methods working together, and propose a feedback mechanism that allows the admission control threshold to be tuned to maximise the efficient usage of available bandwidth in the network, while ensuring that the QoS requirements of all realtime flows are met. We also examine heterogeneous/vertical handover, providing an overview of the technologies supporting seamless handover. The issues studied in this area include a method of using the Signal to Noise Ratio to trigger handover in heterogeneous networks and QoS Mapping between heterogeneous networks. Our proposed method of QoS mapping establishes the minimum set of QoS parameters applicable to individual flows, and then maps these parameters into system parameter formats for both 802.11e and 802.16e networks.

Submitted by Ana Caroline Costa (ana_caroline212@hotmail.com) on 2018-11-09T18:31:42Z No. of bitstreams: 2 Dissertação - Bruno Soares da Silva - 2018.pdf: 2828057 bytes, checksum: dd61488139898a6d0bf95f84ee3d5ddf (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-11-12T13:12:52Z (GMT) No. of bitstreams: 2 Dissertação - Bruno Soares da Silva - 2018.pdf: 2828057 bytes, checksum: dd61488139898a6d0bf95f84ee3d5ddf (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-11-12T13:12:52Z (GMT). No. of bitstreams: 2 Dissertação - Bruno Soares da Silva - 2018.pdf: 2828057 bytes, checksum: dd61488139898a6d0bf95f84ee3d5ddf (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-10-08 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The movement flow detection in indoor environments requires the aquisition and implantation of specialized devices. The perturbations that can affect the electromagnetic signals used by 802.11 interfaces make this type of device a low-cost and widely available movement sensor. Most indoor environments have a 802.11 interface, which makes the use of this type of devices a good option as it doesn't requires any new device. In this work, we propose the WiDMove, a proposal to detect the movement flows in an indoor environment using the channel quality measurements (known as Channel State Information - CSI) offered by the IEEE 802.11n standard. Our proposal is based on signal processing and pattern recognition techniques, which allow us to extract and classify event signatures using the CSI. In lab tests with off-the-shelf 802.11 interfaces, we collected CSI samples that were affected by 8 different people. From this collected data we extracted the signature of the entry and exit events using some techniques such as Principal Component Analysis (PCA), Short-Time Fourier Transform (STFT) and Continuous Wavelet Transform (CWT). We trained two model types, the first based on a Support Vector Machine (SVM) classifier and the second based on a Multi Layer Perceptral (MLP) neural network. We validated this models with average accuracy experiments and with the cross-validation, including the K-Fold and Leave-One-Out techniques. WiDMove presented that can reach an average accuracy above 93% and that we can train neural networks that can reach an accuracy above 97%. / A detecção de fluxos de movimento em ambientes fechados demanda a aquisição e implantação de sensores especializados. As perturbações que podem afetar o sinal eletromagnético utilizado por interfaces de rede 802.11 tornam esse tipo de dispositivo um sensor de fluxos de movimento de baixo custo e amplamente disponível. Grande parte dos ambientes possuem interfaces 802.11 implantadas, tornando o uso desse tipo de dispositivo como sensor de fluxo de movimento bastante viável por não demandar a aquisição de nenhum novo equipamento. Neste trabalho, apresentamos o WiDMove, uma proposta que visa detectar fluxos de movimento em ambientes fechados utilizando as medidas de qualidade do canal oferecidas pelo padrão IEEE 802.11n, conhecidas como Channel State Information (CSI). Nossa proposta é baseada em técnicas de processamento de sinais e de reconhecimento de padrões, as quais nos permitem extrair e classificar assinaturas de fluxos de movimento usando as medidas CSI. Em testes de laboratório com interfaces 802.11 convencionais, coletamos medidas CSI influenciadas por 8 indivíduos distintos e extraímos as assinaturas de entrada e saída utilizando, dentre outras técnicas, Principal Component Analysis (PCA), Short- Time Fourier Transform (STFT) e Continuous Wavelet Transform (CWT). Treinamos dois tipos de modelos, um baseado em um classificador do tipo Support Vector Machine (SVM) e outro baseado em redes neurais do tipo Multi Layer Perceptral (MLP). Validamos esses modelos através de testes de acurácia média e de técnicas de validação cruzada, incluindo as técnicas K-Fold e Leave-One-Out. Os testes demonstraram que o WiDMove pode atingir uma acurácia média superior a 93% e que é possível treinar uma rede neural com acurácia de cerca de 97%.

Our work focuses on wireless networks in general, but deals specifically with security in wireless sensor networks and energy consumption in IEEE 802.11 infrastructure WLANs. In the first part of our work, we focus on secure communication among sensor nodes in a wireless sensor network. These networks consists of large numbers of devices having limited energy and memory. Public key cryptography is too demanding for these resource-constrained devices because it requires high computation. So, we focus on symmetric key cryptography to achieve secure communication among nodes. For this cryptographic technique to work, two nodes have to agree upon a common key. To achieve this, many key distribution schemes have been proposed in the literature. Recently, several researchers have proposed schemes in which they have used group-based deployment models and assumed predeployment knowledge of the expected locations of nodes. They have shown that these schemes achieve better performance than the earlier schemes, in terms of connectivity, resilience against node capture and storage requirements. But in many situations expected locations of nodes are not available. We propose a solution which does not use the group-based deployment model and predeployment knowledge of the locations of nodes, and yet performs better than schemes which make the aforementioned assumptions. In our scheme, groups are formed after the deployment of sensor nodes on the basis of their physical locations. Nodes in different groups sample keys from disjoint key pools, so that compromise of a node affects secure links of its group only. Because of this reason, our scheme performs better than earlier schemes as well as the schemes using predeployment knowledge, in terms of connectivity, storage requirement, and security. Moreover, the post-deployment key generation process completes sooner than in schemes like LEAP+. In the second part of our work, we develop analytical models for estimating the energy spent by stations (STAs) in infrastructure WLANs when performing TCP-controlled file downloads. We focus on the energy spent in radio communication when the STAs are in the Continuously Active Mode (CAM), or in the static Power Save Mode (PSM). Our approach is to develop accurate models for obtaining the fractions of times the STA radios spend in idling, receiving and transmitting. We discuss two traffic models for each mode of operation: (i) each STA performs one large file download, and (ii) the STAs perform short file transfers with think times (short duration of inactivity)between two transfers. We evaluate the rate of STA energy expenditure with long file downloads, and show that static PSM is worse than using just CAM. For short file downloads, we compute the number of file downloads that can be completed with a given battery capacity, and show that PSM performs better than CAM for this case. We provide a validation of our analytical models using the NS-2 simulator. Although the PSM performs better than the CAM when the STAs download short files over TCP with think times, its performance degrades as the number of STAs associated to the access point (AP) increases. To address this problem, we propose an algorithm, which we call opportunistic PSM (OPSM). We show through simulations that OPSM performs better than PSM. The performance gain achieved by OPSM increases as the file size requested by the STAs or the number of STAs associated with the AP increases. We implemented OPSM in NS-2.33, and to compare the performance of OPSM and PSM, we evaluate the number of file downloads that can be completed with a given battery capacity and the average time taken to download a file.

This thesis is concerned with analytical modeling of Wireless Local Area Networks (WLANs) that are based on IEEE 802.11 Distributed Coordination Function (DCF). Such networks are popularly known as WiFi networks. We have developed accurate analytical models for the following three network scenarios: (S1) A single cell WLAN with homogeneous nodes and Poisson packet arrivals, (S2) A multi-cell WLAN (a) with saturated nodes, or (b) with TCP-controlled long-lived downloads, and (S3) A multi-cell WLAN with TCP-controlled short-lived downloads. Our analytical models are simple Markovian abstractions that capture the detailed network behavior in the considered scenarios. The insights provided by our analytical models led to two applications: (i) a faster “model-based'” simulator, and (ii) a distributed channel assignment algorithm. We also study the stability of the network through our Markov models. For scenario (S1), we develop a new approach as compared to the existing literature. We apply a “State Dependent Attempt Rate'” (SDAR) approximation to reduce a single cell WLAN with non-saturated nodes to a coupled queue system. We provide a sufficient condition under which the joint queue length Markov chain is positive recurrent. For the case when the arrival rates into the queues are equal we propose a technique to reduce the state space of the coupled queue system. In addition, when the buffer size of the queues are finite and equal we propose an iterative method to estimate the stationary distribution of the reduced state process. Our iterative method yields accurate predictions for important performance measures, namely, “throughput'”, “collision probability” and “packet delay”. We replace the detailed implementation of the MAC layer in NS-2 with the SDAR contention model, thus yielding a ``model-based'' simulator at the MAC layer. We demonstrate that the SDAR model of contention provides an accurate model for the detailed CSMA/CA protocol in scenario (S1). In addition, since the SDAR model removes much of the details at the MAC layer we obtain speed-ups of 1.55-5.4 depending on the arrival rates and the number of nodes in the single cell WLAN. For scenario (S2), we consider a restricted network setting where a so-called “Pairwise Binary Dependence” (PBD) condition holds. We develop a first-cut scalable “cell-level” model by applying the PBD condition. Unlike a node- or link-level model, the complexity of our cell-level model increases with the number of cells rather than with the number of nodes/links. We demonstrate the accuracy of our cell-level model via NS-2 simulations. We show that, as the “access intensity” of every cell goes to infinity the aggregate network throughput is maximized. This remarkable property of CSMA, namely, “maximization of aggregate network throughput in a distributed manner” has been proved recently by Durvy et al. (TIT, March, 2009) for an infinite linear chain of nodes. We prove it for multi-cell WLANs with arbitrary cell topology (under the PBD condition). Based on this insight provided by our analytical model we propose a distributed channel assignment algorithm. For scenario (S3), we consider the same restricted network setting as for scenario (S2). For Poisson flow arrivals and i.i.d. exponentially distributed flow sizes we model a multi-cell WLAN as a network of processor-sharing queues with state-dependent service rates. The state-dependent service rates are obtained by applying the model for scenario (S2) and taking the access intensities to infinity. We demonstrate the accuracy of our model via NS-2 simulations. We also demonstrate the inaccuracy of the service model proposed in the recent work by Bonald et al. (SIGMETRICS 2008) and identify the implicit assumption in their model which leads to this inaccuracy. We call our service model which accurately characterizes the service process in a multi-cell WLAN (under the PBD condition) “DCF scheduling” and study the “stability region” of DCF scheduling for small networks with single or multiple overlapping “contention domains”.