No description available.

Made available in DSpace on 2016-08-29T15:33:19Z (GMT). No. of bitstreams: 1 tese_7981_Dissertação - Vinicius Barcellos Antunes20150602-133738.pdf: 1742369 bytes, checksum: e6b03688b02047f61c0bffaa0fdb7c4c (MD5) Previous issue date: 2014-08-29 / A pesquisa na área de Redes de Sensores sem Fio (RSSF) tem contribuído de forma significativa para o desenvolvimento de aplicações de sensoriamento em larga escala. Em função de suas características e restrições (pouca capacidade de processamento, armazenamento e fonte de energia), diferentes protocolos de comunicação foram projetados para as RSSFs. Para atender a uma necessidade de padronização, o IETF (Internet Engineering Task Force) especificou um protocolo de roteamento denominado RPL (IPv6 Routing Protocol for Low-power and Lossy Networks), como protocolo padrão para redes RSSFs. Este protocolo possui uma grande quantidade de recursos e também fornece facilidades para incorporar métricas de natureza dinâmica.De forma a facilitar o acesso a uma variedade de informações requeridas pelo protocolo de roteamento adaptativo RPL para cálculo métricas e tomadas de decisão de encaminhamento, é proposto um gerenciador de contexto reconfigurável. Este gerenciador é responsável por coletar os dados contextuais das fontes de informação disponíveis, e atuar como uma camada de abstração, simplificando o uso dessas fontes e habilitando a experimentação e prototipagem de novas métricas de roteamento.Esta dissertação apresenta uma extensão ao protocolo de roteamento RPL. Esta extensão tem por objetivo prover mecanismos simples para que o protocolo RPL interaja com a aplicação, com a finalidade de otimização do roteamento. Tal interação é realizada através de regras de aplicação, permitindo ao RPL se adaptar em tempo de execução a mudanças no ambiente.

There has been a movement across South Africa's education sector to widen access and participation in higher education (HE). Recognition of Prior Learning (RPL) programmes offer an opportunity to do so by providing access to HE to those who do not necessarily have the prerequisite requirements. There is still much to learn regarding RPL programmes and the unique needs and challenges of RPL students, as most of these programmes are still in the pilot stage. This study is specifically concerned with the transition of RPL students into postgraduate HE, as little research has been conducted on this topic. To gain a deeper understanding of RPL student transition, the researcher followed a case-study approach in which the experiences of RPL students enrolled in an online postgraduate diploma programme in management in marketing (PgDipMM) at a South African university were investigated. The programme piloted a novel RPL assessment and selection process which aimed to address criticisms of previous models. Qualitative data was obtained through eight open-ended, in-depth interviews with RPL students enrolled in the programme. Using a thematic analysis approach, nine key themes were identified as possible mechanism that facilitated RPL students' successful transition into the postgraduate diploma and six themes that did not. Findings indicated that, to a large extent, the novel RPL assessment and selection process appeared to have facilitated RPL students' successful transition into a postgraduate diploma through technical preparation and building students' confidence in their abilities. Possible enablers and barriers to RPL students' successful transition were also identified and discussed. In short, the enablers included learner maturity, transference of prior knowledge and skills, social support, sense of belonging, access to effective resources and financial security. Conversely, the barriers included academic difficulties, under preparedness for the intensity of HE and time restrictions. The study had theoretical and practical implications in that it contributed to our understanding of RPL student transition to HE and provided suggestions for ways in which their successful transition can be facilitated

Energy is an important factor that must be considered by multi-hop wireless mesh routing protocols because most sensors are powered by batteries with a limited capacity. We focus on the industry-standard RPL (Routing Protocol over Low-power and lossy networks) routing protocol that must find energy-efficient paths in low-power and lossy networks. However, the existing RPL objective functions route based on hop-count and ETX (expected transmission count) metrics alone, ignoring the energy cost of data transmission and reception. We address this issue in two ways. First, we design an objective function for RPL that finds paths that require, in expectation, the minimum amount of energy. Second, we design a probing mechanism which configures the transmission power of sensors to minimize energy consumption. The proposed approach is implemented and evaluated using simulations as well as on a small testbed with two Zolertial Z1 motes.

The main aim of this dissertation is to address the security issues of the physical layer of Cyber Physical Systems. The network security is first assessed using a 5-level Network Security Evaluation Scheme (NSES). The network security is then enhanced using a novel Intrusion Detection System that is designed using Supervised Machine Learning. Defined as a complete architecture, this framework includes a complete packet analysis of radio traffic of Routing Protocol for Low-Power and Lossy Networks (RPL). A dataset of 300 different simulations of RPL network is defined for normal traffic, hello flood attack, DIS attack, increased version attack and decreased rank attack. The IDS is a multi-model detection model that provides an efficient detection against the known as well as new attacks. The model analysis is done with the cross-validation method as well as using the new data from a similar network. To detect the known attacks, the model performed at 99% accuracy rate and for the new attack, 85% accuracy is achieved. / Graduate

L'intérêt grandissant pour l'Internet des Objets (IdO) s'est traduit par le déploiement à grande échelle de réseaux dits LLN. Ceux-ci sont fortement contraints en matière de ressources et communiquent via des liens instables. Les protocoles de routages existants pour les réseaux traditionnels ne sont pas adaptés à ces caractéristiques. L'IETF a proposé un nouveau protocole de routage appelé RPL fondé sur IPv6 et spécifiquement conçu pour ces environnements. Cependant, il est exposé à de nombreuses attaques et la mise en place de mécanismes de sécurité représente un coût considérable. Les réseaux LLN introduisent donc de nouveaux enjeux quant à leur supervision et leur sécurité. Dans le cadre de cette thèse, nous étudions une approche de supervision pour la sécurité de l'IdO tout en limitant son coût. Nous évaluons tout d'abord les menaces auxquelles sont soumis les réseaux RPL en les classifiant au travers d'une taxonomie. Nous quantifions également les conséquences de deux attaques appelées l'attaque d'incohérence DAG et l'attaque du numéro de version. Nous nous concentrons ensuite sur les solutions pour la sécurité dans les réseaux RPL. Nous proposons une stratégie locale qui limite les attaques d'incohérences DAG et nous présentons une architecture de supervision distribuée orientée sécurité pour détecter des attaques complexes comme les attaques sur le numéro de version et compléter notre approche locale. Celle-ci nous permet de préserver l'énergie des nœuds en effectuant les activités de surveillance et de détection sur des nœuds dédiés. Nous quantifions ensuite les performances de cette architecture ainsi que la stratégie de détection proposée. / The growing interest for the Internet of Things (IoT) has resulted in the large scale deployment of Low power and Lossy Networks (LLN). These networks are strongly constrained in terms of resources and communicate using unstable links. In this context, existing routing protocols for traditional networks do not cope with all these constraints. The IETF has proposed a new routing protocol called RPL based on IPv6 and specifically designed for these environments. The RPL protocol is however exposed to a large variety of attacks. The deployment of security mechanisms may also be quite expensive for the nodes. Therefore, LLN networks present new challenges in terms of monitoring and security. In this thesis we propose to investigate a security-oriented monitoring approach for addressing the trade-off between security and cost in the IoT. In a first stage, we assess security threats faced by these networks by identifying and classifying attacks through a dedicated taxonomy. We also quantify the consequences of two major attacks called DAG inconsistency attacks and version number attacks causing over-consumption of node resources. We then focus our work on security solutions for RPL-based IoT. We propose a local strategy for addressing DAG inconsistency attacks. In order to detect complex attacks such as version number attacks and to complement our node-level approach, we design a security-oriented distributed monitoring architecture for RPL networks. This solution allows us to preserve constrained nodes energy by performing monitoring and detection activities on dedicated nodes. We quantify the performance and the cost of this architecture and the deployed detection modules.

Faculty of Humanities School of Education 0415279k jthakrar@ofh.ac.za / Recognition of prior learning (RPL), the practice of valuing adults’ learning, derived from formal, informal or non-formal contexts, forms a key component of the transformational agenda of the education landscape in South Africa. Yet few institutions of higher education in South Africa practice RPL. The aim of this study is to provide insight into RPL by evaluating its current practice at one institution of higher education, the University of Fort Hare, highlighting its successes and challenges. The sources of information for this study included past RPL candidates, RPL assessors and co-ordinators, and the policy, processes and procedures that govern the RPL process in the institution. The methods used to collect the data included interviews, observation of institutional RPL workshops and an analysis of relevant RPL documentation. The research results suggest that there has been a limited uptake of RPL, which has impeded a holistic approach to RPL. Challenges include the draft status of an institutional policy, inconsistent RPL practices, the limitations of some of the RPL assessment methods, the lack of RPL data systems, the costs associated with supporting RPL candidates, the duplication of resources and practices, and the impact on academic workloads. A series of recommendations have been made to address these and other challenges highlighted in the research results.

Les Smart Grids visent à transformer le réseau électrique actuel en un réseau "plus intelligent" où la production énergétique est décentralisée et automatisée, facilitant l'intégration des sources d'énergie renouvelables. Cette évolution est rendue possible grâce à l'utilisation d'un réseau de communication pour les multiples échanges de données hétérogènes des Smart Grids. L'objectif de cette thèse est de proposer un paradigme de communication efficace en termes de qualité de service pour les Smart Grids basé sur les réseaux de capteurs.Dans un premier temps, on s’intéresse au protocole standard RPL. Nous proposons une évolution de celui-ci à travers une nouvelle fonction objectif. Celle-ci tire parti de l’hétérogénéité matérielle des nœuds et des liens pour introduire la qualité de service. Cela permet à RPL de satisfaire les multiples et différentes exigences en termes de fiabilité, de latence et de priorité dans l'acheminement des données. Nos résultats montrent que notre approche permet bien la différentiation du trafic tout en réduisant la latence du routage et en économisant l'énergie.Nous proposons également d'améliorer l'utilisation du réseau de capteurs en y introduisant l’auto-organisation et la réduction des données. Le but est alors de prédire la valeur des données mesurées plutôt que de les transmettre.Une autre approche explorée est d'agréger les différents messages transitant sur le réseau tout en considérant leurs différentes exigences de qualité de service. Ces deux approches permettent ainsi de réduire la consommation d'énergie tout en respectant les exigences des différentes applications des Smart Grids. / Smart Grids aim to transform the current electric grid into a "smarter" network where energy production is decentralized and automated, which facilitates the integration of renewable energy resources. This evolution is made possible thanks to the use of a communication network for the multiple heterogeneous data exchanges of the Smart Grids. Hence, the aim of this thesis is to propose an efficient communication paradigm in terms of quality of service for Smart Grids based on wireless sensor networks.First, we study data routing in Smart Grids with the RPL standard. Nevertheless, RPL is not suitable for Smart Grid applications in terms of quality of service. Therefore, we propose an objective function for RPL that takes different features of both nodes and links into consideration. Results show that our approach improves network performance compared to existing solutions in terms of packet delivery ratio, network lifetime, latency and traffic differentiation.Then, we also propose a more efficient data collection by introducing self-organization and data reduction for these wireless sensors. The goal is to predict the value of the measured data rather than transmitting them. Another explored approach is to aggregate the different messages sent across the network while considering their different requirements in terms of quality of service.These two approaches reduce the energy consumption while respecting the requirements of the different applications of the Smart Grids.

Low-Power Lossy Networks (LLNs) are a type of Internet of Things (IoT) meshnetwork that collaboratively interact and perform various tasks autonomously. TheRouting Protocol for Low-power and Lossy Network (RPL) is the most used rout-ing protocol for LLNs. Recently, we have been witnessing a tremendous increasein attacks on Internet infrastructures using IoT devices as a botnet (IoT botnet).This thesis focuses on two parts: designing an ML-based IDS for 6LoWPAN, andgenerating a new larger labeled RPL attack dataset by implementing various non-attack and attack IoT network scenarios in the Cooja simulator. The collected rawdata from simulations is preprocessed and labeled to train the Machine Learningmodel for Intrusion Detection System (IDS). We used Deep Neural Network (DNN),Random Forest Classifier (RFC), and Support Vector Machines with Radial-BasisFunction kernel (SVM-RBF) learning algorithms to detect attack in RPL based IoTmesh networks. We achieved a high accuracy (96.7%) and precision (95.7%) usingthe RFC model. The thesis also reviewed the possible placement strategy of IDSfrom cloud to edge.

Le réseau électrique a subi d’importantes évolutions ces dernières décennies, pour devenir ce qu’on appelle le Smart Grid. Le réseau électrique évolue actuellement d’une architecture centralisée vers une architecture décentralisée, tenant compte des consommations et sources d’énergies à caractère imprédictible et irrégulier. L’Advanced Metering Infrastructure est une architecture clé du Smart Grid qui permet des communications bidirectionnelles entre le consommateur et le fournisseur d’énergie. Les réseaux de compteurs intelligents qui constituent notamment cette architecture reposent communément sur des communications à courant porteur, une technologie qui est hautement sensible aux interférences. Malgré l’utilisation de protocoles de niveau 2 spécifiques, les technologies employées ne permettent pas de respecter les exigences de toutes les applications Smart grid. La plupart des technologies considérées pour les réseaux de compteurs intelligents sont de courte portée, chaque compteur ne peut communiquer directement avec le concentrateur. Les noeuds doivent collaborer entre eux, utilisant un protocole de routage tel que RPL pour atteindre la destination. Le but de cette thèse est d’adapter RPL à un environnement multi interfaces, et étudier comment l’hétérogénéité des interfaces peut améliorer la fiabilité et les performances d’un réseau de compteurs intelligents. / Since decades, the power grid is Under going a tremendous evolution, toward what is called the Smart Grid. The grid is actually evolving from a centralized architecture to a decentralized one, taking into account all the unpredictable sources and consumption. The Advanced Metering Infrastructure is the network dedicated to the Smart Grid that allows two-ways communications between the consumers and the energy providers. Smart Meters networks, that are part of this architecture, rely on powerline communications, a technology that is highly sensitive to interference. Despite dedicated layer 2 protocols, the employed technologies cannot fulfill most of smart grid applications requirements. The majority of smart meter network technology candidates are short range, each meter cannot reach the concentrator in one hop. Nodes need to collaborate, using a routing protocol like RPL to reach the destination. The goal of this thesis is to modify RPL to a multi interfaces environment, and study how interface heterogeneity could increase the reliability and the performance of a smart meter network.