<p> <p>&nbsp / </p> </p> <p align="left">The underlying theme of this thesis is identification, classification, detection and prediction of cellular network faults using state of the art technologies, methods and algorithms.</p>

In mobile and wireless networks, data broadcasting for popular data items enables the efficient utilization of the limited wireless bandwidth. However, efficient data scheduling schemes are needed to fully exploit the benefits of data broadcasting. Towards this goal, several broadcast scheduling policies have been proposed. These existing schemes have mostly focused on either minimizing response time, or drop rate, when requests are associated with hard deadlines. The inherent inaccuracy of hard deadlines in a dynamic mobile environment motivated us to use Service Level Agreements (SLAs) where a user specifies the utility of data as a function of its arrival time. Moreover, SLAs provide the mobile user with an already familiar quality of service specification from wired environments. Hence, in this dissertation, we propose SAAB, an SLA-aware adaptive data broadcast scheduling policy for maximizing the system utility under SLA-based performance measures. To achieve this goal, SAAB considers both the characteristics of disseminated data objects as well as the SLAs associated with them. Additionally, SAAB automatically adjusts to the system workload conditions which enables it to constantly outperform existing broadcast scheduling policies.

In mobile and wireless networks, data broadcasting for popular data items enables the efficient utilization of the limited wireless bandwidth. However, efficient data scheduling schemes are needed to fully exploit the benefits of data broadcasting. Towards this goal, several broadcast scheduling policies have been proposed. These existing schemes have mostly focused on either minimizing response time, or drop rate, when requests are associated with hard deadlines. The inherent inaccuracy of hard deadlines in a dynamic mobile environment motivated us to use Service Level Agreements (SLAs) where a user specifies the utility of data as a function of its arrival time. Moreover, SLAs provide the mobile user with an already familiar quality of service specification from wired environments. Hence, in this dissertation, we propose SAAB, an SLA-aware adaptive data broadcast scheduling policy for maximizing the system utility under SLA-based performance measures. To achieve this goal, SAAB considers both the characteristics of disseminated data objects as well as the SLAs associated with them. Additionally, SAAB automatically adjusts to the system workload conditions which enables it to constantly outperform existing broadcast scheduling policies.

No description available.

<p> <p>&nbsp / </p> </p> <p align="left">The underlying theme of this thesis is identification, classification, detection and prediction of cellular network faults using state of the art technologies, methods and algorithms.</p>

Les centres de données verts de dernière génération ont été conçus pour une consommation optimisée et une meilleure qualité du niveau de service SLA. Cependant,ces dernières années, le marché des centres de données augmente rapidement,et la concentration de la puissance de calcul est de plus en plus importante, ce qui fait augmenter les besoins en puissance électrique et refroidissement. Un centre de données est constitué de ressources informatiques, de systèmes de refroidissement et de distribution électrique. De nombreux travaux de recherche se sont intéressés à la réduction de la consommation des centres de données afin d'améliorer le PUE, tout en garantissant le même niveau de service. Certains travaux visent le dimensionnement dynamique des ressources en fonction de la charge afin de réduire le nombre de serveurs démarrés, d'autres cherchent à optimiser le système de refroidissement qui représente un part important de la consommation globale.Dans cette thèse, afin de réduire le PUE, nous étudions la mise en place d'un système autonome d'optimisation globale du refroidissement, qui se base sur des sources de données externes tel que la température extérieure et les prévisions météorologiques, couplé à un module de prédiction de charge informatique globale pour absorber les pics d'activité, pour optimiser les ressources utilisés à un moindre coût, tout en préservant la qualité de service. Afin de garantir un meilleur SLA, nous proposons une architecture distribuée pour déceler les anomalies de fonctionnements complexes en temps réel, en analysant de gros volumes de données provenant des milliers de capteurs du centre de données. Détecter les comportements anormaux au plus tôt, permet de réagir plus vite face aux menaces qui peuvent impacter la qualité de service, avec des boucles de contrôle autonomes qui automatisent l'administration. Nous évaluons les performances de nos contributions sur des données provenant d'un centre de donnée en exploitation hébergeant des applications réelles. / Next-generation green datacenters were designed for optimized consumption and improved quality of service level Service Level Agreement (SLA). However, in recent years, the datacenter market is growing rapidly, and the concentration of the computing power is increasingly important, thereby increasing the electrical power and cooling consumptions. A datacenter consists of computing resources, cooling systems, and power distribution. Many research studies have focused on reducing the consumption of datacenters to improve the PUE, while guaranteeing the same level of service. Some works aims the dynamic sizing of resources according to the load, to reduce the number of started servers, others seek to optimize the cooling system which represents an important part of total consumption. In this thesis, in order to reduce the PUE, we study the design of an autonomous system for global cooling optimization, which is based on external data sources such as the outside temperature and weather forecasting, coupled with an overall IT load prediction module to absorb the peaks of activity, to optimize activere sources at a lower cost while preserving service level quality. To ensure a better SLA, we propose a distributed architecture to detect the complex operation anomalies in real time, by analyzing large data volumes from thousands of sensors deployed in the datacenter. Early identification of abnormal behaviors, allows a better reactivity to deal with threats that may impact the quality of service, with autonomous control loops that automate the administration. We evaluate the performance of our contributions on data collected from an operating datacenter hosting real applications.

Le Cloud Computing est en plein essor, grace a ses divers avantages, telsl'elasticite, le cout, ou encore son importante exibilite dans le developpementd'applications. Il demeure cependant des problemes en suspens, lies auxperformances, a la disponibilite, la fiabilite, ou encore la securite. De nombreusesetudes se focalisent sur la fiabilite et les performance dans les servicesdu Cloud, qui sont les points critiques pour le client. On retrouve parmicelles-ci plusieurs themes emergents, allant de l'ordonnancement de tachesau placement de donnees et leur replication, en passant par la tolerance auxfautes adaptative ou a la demande, et l'elaboration de nouveaux modeles defautes.Les outils actuels evaluant la fiabilite des services du Cloud se basent surdes parametres simplifies. Ils ne permettent pas d'analyser les performancesou de comparer l'efficacite des solutions proposees. Cette these aborde precisement ce probleme en proposant un modele d'environnement complet detest destine a evaluer la fiabilite et les performances des services de CloudComputing. La creation d'outils de tests destines a l'evaluation de la fiabiliteet des performances des services du Cloud pose de nombreux defis, en raisonde la grande quantite et de la complexite des donnees traitees par ce genrede services. Les trois principaux modeles de Cloud Computing sont respectivement:Infrastructure en tant que Service (IaaS), Plate-forme en tant queService (PaaS) et Logiciel en tant que Service (SaaS).Dans le cadre de cettethese, nous nous concentrons sur le modele PaaS. Il permet aux systemesd'exploitation ou aux intergiciels d'etre accessibles via une connexion internet.Nous introduisons une architecture de test generique, qui sera utiliseepar la suite lors de la creation d'outils de tests, destines a l'evaluation de lafiabilite et de la performance. / Cloud computing models are attractive because of various benefits such asscalability, cost and exibility to develop new software applications. However,availability, reliability, performance and security challenges are still not fullyaddressed. Dependability is an important issue for the customers of cloudcomputing who want to have guarantees in terms of reliability and availability.Many studies investigated the dependability and performance of cloud services,ranging from job scheduling to data placement and replication, adaptiveand on-demand fault-tolerance to new fault-tolerance models. However, thead-hoc and overly simplified settings used to evaluate most cloud service fault toleranceand performance improvement solutions pose significant challengesto the analysis and comparison of the effectiveness of these solutions.This thesis precisely addresses this problem and presents a benchmarkingapproach for evaluating the dependability and performance of cloud services.Designing of dependability and performance benchmarks for a cloud serviceis a particular challenge because of high complexity and the large amount ofdata processed by such service. Infrastructure as a Service (IaaS), Platform asa Service (PaaS) and Software as a Service (SaaS) are the three well definedmodels of cloud computing. In this thesis, we will focus on the PaaS modelof cloud computing. PaaS model enables operating systems and middlewareservices to be delivered from a managed source over a network. We introduce ageneric benchmarking architecture which is further used to build dependabilityand performance benchmarks for PaaS model of cloud services.

With Cloud Computing becoming a popular term on the Information Technology (IT) market, security and accountability has become important issues to highlight. In our research we review these concepts by focusing on security risks with Cloud Computing and the associated services; Software, Platform and Infrastructure (SPI) and connecting them with a social study of trust. The method that was conducted during our research was reviewing secondary literature, interviewing different experts regarding Cloud Computing and relating standards already established by ENISA, NIST, and CSA to the interviews. The result of this study shows connections between the specific SPIs, both how they compare, but also how they differ. In the end we were also able to rank the top security risks from interviews with experts and see which SPI could be the most insecure one and  what countermeasures could be applied. This was further related to trust and Service Level Agreement (SLA) in Cloud Computing to show how the security risks we discuss are related to these two specific areas. By highlighting this we wanted to present useable information for both clients and providers in how to create a better Cloud Computing environment.

Philosophiae Doctor - PhD / The underlying theme of this thesis is identification, classification, detection and prediction of cellular network faults using state of the art technologies, methods and algorithms.

An important goal of modern data networks is to support multiple applications over a single network infrastructure. The combination of data, voice, video and conference traffic, each requiring a unique Quality of Service (QoS), makes resource dimensioning a very challenging task. To guarantee QoS by mere over-provisioning of bandwidth is not viable in the long run, as network resources are expensive. The aim of proper resource dimensioning is to provide the required QoS while making optimal use of the allocated bandwidth. Dimensioning parameters used by service providers today are based on best practice recommendations, and are not necessarily optimal. This dissertation focuses on resource dimensioning for the DiffServ network architecture. Four predefined traffic classes, i.e. Real Time (RT), Interactive Business (IB), Bulk Business (BB) and General Data (GD), needed to be dimensioned in terms of bandwidth allocation and traffic regulation. To perform this task, a study was made of the DiffServ mechanism and the QoS requirements of each class. Traffic generators were required for each class to perform simulations. Our investigations show that the dominating Transport Layer protocol for the RT class is UDP, while TCP is mostly used by the other classes. This led to a separate analysis and requirement for traffic models for UDP and TCP traffic. Analysis of real-world data shows that modern network traffic is characterized by long-range dependency, self-similarity and a very bursty nature. Our evaluation of various traffic models indicates that the Multi-fractal Wavelet Model (MWM) is best for TCP due to its ability to capture long-range dependency and self-similarity. The Markov Modulated Poisson Process (MMPP) is able to model occasional long OFF-periods and burstiness present in UDP traffic. Hence, these two models were used in simulations. A test bed was implemented to evaluate performance of the four traffic classes defined in DiffServ. Traffic was sent through the test bed, while delay and loss was measured. For single class simulations, dimensioning values were obtained while conforming to the QoS specifications. Multi-class simulations investigated the effects of statistical multiplexing on the obtained values. Simulation results for various numerical provisioning factors (PF) were obtained. These factors are used to determine the link data rate as a function of the required average bandwidth and QoS. The use of class-based differentiation for QoS showed that strict delay and loss bounds can be guaranteed, even in the presence of very high (up to 90%) bandwidth utilization. Simulation results showed small deviations from best practice recommendation PF values: A value of 4 is currently used for both RT and IB classes, while 2 is used for the BB class. This dissertation indicates that 3.89 for RT, 3.81 for IB and 2.48 for BB achieve the prescribed QoS more accurately. It was concluded that either the bandwidth distribution among classes, or quality guarantees for the BB class should be adjusted since the RT and IB classes over-performed while BB under-performed. The results contribute to the process of resource dimensioning by adding value to dimensioning parameters through simulation rather than mere intuition or educated guessing. / Dissertation (MEng (Electronic Engineering))--University of Pretoria, 2007. / Electrical, Electronic and Computer Engineering / unrestricted