L’informatique, le stockage des données et les applications à la demande font partie des services offerts par l’architecture informatique en Nuage. Dans ce cadre, les fournisseurs de nuage (FN) agissent non seulement en tant qu’administrateurs des ressources d'infrastructure mais ils profitent aussi financièrement de la location de ces ressources. Dans cette thèse, nous proposons trois modèles d'optimisation du processus d'allocation des ressources dans le nuage dans le but de réduire les coûts générés et d’accroitre la qualité du service rendu. Cela peut être accompli en fournissant au FN les outils formels nécessaires pour réduire au minimum le prix des ressources dédiées à servir les requêtes des utilisateurs. Ainsi, la mise en œuvre des modèles proposés permettra non seulement l’augmentation des revenus du FN, mais aussi l’amélioration de la qualité des services offerts, ce qui enrichira l’ensemble des interactions qui se produisent dans le nuage. A cet effet, nous nous concentrons principalement sur les ressources de l’infrastructure en tant que service (IaaS), lesquels sont contenus dans des centres de données (DCs), et constituent l'infrastructure physique du nuage. Comme une alternative aux immenses DCs centralisés, la recherche dans ce domaine comprend l’installation de petits centres de données (Edge DCs) placés à proximité des utilisateurs finaux. Dans ce contexte nous adressons le problème d’allocation des ressources et pour ce faire nous utilisons la technique d'optimisation nommée génération de colonnes. Cette technique nous permet de traiter des modèles d'optimisation à grande échelle de manière efficace. La formulation proposée comprend à la fois, et dans une seule phase, les communications et les ressources informatiques à optimiser dans le but de servir les requêtes de service d'infrastructure. Sur la base de cette formulation, nous proposons également un deuxième modèle qui comprend des garanties de qualité de service toujours sous la même perspective d'allocation des ressources d’infrastructure en tant que service. Ceci nous permet de fournir plusieurs solutions applicables à divers aspects du même problème, tels que le coût et la réduction des délais, tout en offrant différents niveaux de service. En outre, nous introduisons le scénario informatique en nuage multimédia, qui, conjointement avec l'architecture des Edge DCs, résulte en l'architecture Multimédia Edge Cloud (MEC). Dans ce cadre, nous proposons une nouvelle approche pour l'allocation des ressources dans les architectures informatique en nuage multimédia lors du positionnement de ces DCs afin de réduire les problèmes liés à la communication, tels que la latence et la gigue. Dans cette formulation, nous proposons également de mettre en œuvre des technologies optiques de réseau de fibres pour améliorer les communications entre les DCs. Plusieurs travaux ont proposé de nouvelles méthodes pour améliorer la performance et la transmission de données. Dans nos travaux, nous avons décidé de mettre en œuvre le multiplexage en longueur d'onde (WDM) pour renforcer l'utilisation des liens et les chemins optiques dans le but de grouper différents signaux sur la même longueur d'onde. Un environnement de simulation réel est également présenté pour l’évaluation des performances et de l'efficacité des approches proposées. Pour ce faire, nous utilisons le scénario spécifié pour les DCs, et nous comparons par simulation nos modèles au moyen de différents critères de performances tel que l'impact de la formulation optique sur la performance du réseau. Les résultats numériques obtenus ont montré que, en utilisant nos modèles, le FN peut efficacement réduire les coûts d'allocation en maintenant toujours un niveau satisfaisant quant à l'acceptation de requêtes et la qualité du service. / The cloud architecture offers on-demand computing, storage and applications. Within this structure, Cloud Providers (CPs) not only administer infrastructure resources but also directly benefit from leasing them. In this thesis, we propose three optimization models to assist CPs reduce the costs incurred in the resource allocation process when serving users’ demands. Implementing the proposed models will not only increase the CP’s revenue but will also enhance the quality of the services offered, benefiting all parties. We focus on Infrastructure as a Service (IaaS) resources which constitute the physical infrastructure of the cloud and are contained in datacenters (DCs). Following existing research in DC design and cloud computing applications, we propose the implementation of smaller DCs (Edge DCs) be located close to end users as an alternative to large centralized DCs. Lastly, we use the Column Generation optimization technique to handle large scale optimization models efficiently. The proposed formulation optimizes both the communications and information technology resources in a single phase to serve IaaS requests. Based on this formulation, we also propose a second model that includes QoS guarantees under the same Infrastructure as a Service resource allocation perspective, to provide different solutions to diverse aspects of the resource allocation problem such as cost and delay reduction while providing different levels of service. Additionally, we consider the multimedia cloud computing scenario. When Edge DCs architecture is applied to this scenario it results in the creation of the Multimedia Edge Cloud (MEC) architecture. In this context we propose a resource allocation approach to help with the placement of these DCs to reduce communication related problems such as jitter and latency. We also propose the implementation of optical fiber network technologies to enhance communication between DCs. Several studies can be found proposing new methods to improve data transmission and performance. For this study, we decided to implement Wavelength Division Multiplexing (WDM) to strengthen the link usage and light-paths and, by doing so, group different signals over the same wavelength. Using a realistic simulation environment, we evaluate the efficiency of the approaches proposed in this thesis using a scenario specifically designed for the DCs, comparing them with different benchmarks and also simulating the effect of the optical formulation on the network performance. The numerical results obtained show that by using the proposed models, a CP can efficiently reduce allocation costs while maintaining satisfactory request acceptance and QoS ratios.

L’un des principaux défis des fournisseurs de services cloud est d’offrir aux utilisateurs une performance acceptable, tout en minimisant les besoins en matériel et énergie. Dans cette thèse CIFRE menée avec Outscale, un fournisseur de cloud, nous visons à optimiser l’allocation des ressources en utilisant de nouvelles sources d’information. Nous caractérisons la charge de travail pour comprendre le stress résultant sur l’orchestrateur, et la compétition pour les ressources disponibles qui dégrade la qualité de service. Nous proposons un modèle pour prédire la durée d’exécution des VMs à partir de caractéristiques prédictives disponibles au démarrage. Enfin, nous évaluons la sensibilité aux erreurs d’un algorithme de placement des VMs de la littérature qui se base sur ces prédictions. Nous ne trouvons pas d’intérêt à coupler note système prédictif avec cet algorithme, mais nous proposons d’autres façons d’utiliser les prédictions pour optimiser le placement des VMs. / One of the main challenges for cloud computing providers remains to offer trustable performance for all users, while maintaining an efficient use of hardware and energy resources. In the context of this CIFRE thesis lead with Outscale, apublic cloud provider, we perform an in-depth study aimed at making management algorithms use new sources of information. We characterize Outscale’s workload to understand the resulting stress for the orchestrator, and the contention for hardware resources. We propose models to predict the runtime of VMs based on features which are available when they start. We evaluate the sensitivity with respect to prediction error of a VM placement algorithm from the literature that requires such predictions. We do not find any advantage in coupling our prediction model and the selected algorithm, but we propose alternative ways to use predictions to optimize the placement of VMs.

Virtualisation is one of the underlying technologies that led to the success of cloud computing platforms (CCPs). The technology, along with other features such as multitenancy allows delivering of computing resources in the form of service through efficient sharing of physical resources. As these resources are provided through virtualisation, a robust agreement is outlined for both the quantity and quality-of-service (QoS) in a service level agreement (SLA) documents. QoS is one of the essential components of SLA, where performance is one of its primary aspects. As the technology is progressively maturing and receiving massive acceptance, researchers from industry and academia continue to carry out novel theoretical and practical studies of various essential aspects of CCPs with significant levels of success. This thesis starts with the assessment of the current level of knowledge in the literature of cloud computing in general and CCPs in particular. In this context, a substantive literature review was carried out focusing on performance modelling, testing, analysis and evaluation of Infrastructure as a Service (IaaS), methodologies. To this end, a systematic mapping study (SMSs) of the literature was conducted. SMS guided the choice and direction of this research. The SMS was followed by the development of a novel open queueing network model (QNM) at equilibrium for the performance modelling and analysis of an OpenStack IaaS CCP. Moreover, it was assumed that an external arrival pattern is Poisson while the queueing stations provided exponentially distributed service times. Based on Jackson’s theorem, the model was exactly decomposed into individual M/M/c (c ≥ 1) stations. Each of these queueing stations was analysed in isolation, and closed-form expressions for key performance metrics, such as mean response time, throughput, server (resource) utilisation as well as bottleneck device were determined. Moreover, the research was extended with a proposed open QNM with a bursty external arrival pattern represented by a Compound Poisson Process (CPP) with geometrically distributed batches, or equivalently, variable Generalised Exponential (GE) interarrival and service times. Each queueing station had c (c ≥ 1) GE-type servers. Based on a generic maximum entropy (ME) product form approximation, the proposed open GE-type QNM was decomposed into individual GE/GE/c queueing stations with GE-type interarrival and service times. The evaluation of the performance metrics and bottleneck analysis of the QNM were determined, which provided vital insights for the capacity planning of existing CCP architectures as well as the design and development of new ones. The results also revealed, due to a significant impact on the burstiness of interarrival and service time processes, resulted in worst-case performance bounds scenarios, as appropriate. Finally, an investigation was carried out into modelling and analysis of performance and security trade-offs for a CCP architecture, based on a proposed generalised stochastic Petri net (GSPN) model with security-detection control model (SDCM). In this context, ‘optimal’ combined performance and security metrics were defined with both M-type or GE-type arrival and service times and the impact of security incidents on performance was assessed. Typical numerical experiments on the GSPN model were conducted and implemented using the Möbius package, and an ‘optimal’ trade-offs were determined between performance and security, which are crucial in the SLA of the cloud computing services. / Petroleum technology development fund (PTDF) of the government of Nigeria Usmanu Danfodiyo University, Sokoto

Cloud Computing (CC) is one of the fast growing computer network technologies and many companies offer their services through cloud network. Cloud Computing has many properties with respect to the existing traditional service provisions like scalability, availability, fault tolerance, capability and so on which are supported by many IT companies like Google, Amazon, Salesforce.com. These IT companies have more chances to adapt their services into a new environment, known as Cloud computing systems. There are many cloud computing services which are being provided by many IT companies.The purpose of this thesis is to investigate which cloud environment (public, private and hybrid) and services (Infrastructure as a Service, Software as a Service, and Platform as a Service) are suitable for Swedish Armed Forces (SWAF) with respect to performance, security, cost, flexibility and functionality. SWAF is using private (internal) cloud for communications where both sensitive and non-sensitive information are located in the internal cloud. There are problems like maintenance of hardware, cost issues and secure communication while maintaining the private cloud. In order to overcome those problems we have suggested a hybrid and community cloud environment and SaaS, IaaS, PaaS services for SWAF.For suggesting these cloud environments and cloud services we have performed a literature study and two empirical studies (survey and interviews) with different organizations.A new cloud model is designed based on the suggested cloud environment, separate storage spaces for sensitive and non-sensitive information, suitable services and an effective infrastructure for sharing the internal information for SWAF.