Popular fiction books describe rich visual environments that contain characters, objects, and behaviour. This research develops automated processes for converting text sourced from fiction books into animated virtual environments and multi-modal films. This involves the analysis of unrestricted natural language fiction to identify appropriate visual descriptions, and the interpretation of the identified descriptions for constructing animated 3D virtual environments. The goal of the text analysis stage is the creation of annotated fiction text, which identifies visual descriptions in a structured manner. A hierarchical rule-based learning system is created that induces patterns from example annotations provided by a human, and uses these for the creation of additional annotations. Patterns are expressed as tree structures that abstract the input text on different levels according to structural (token, sentence) and syntactic (parts-of-speech, syntactic function) categories. Patterns are generalized using pair-wise merging, where dissimilar sub-trees are replaced with wild-cards. The result is a small set of generalized patterns that are able to create correct annotations. A set of generalized patterns represents a model of an annotator's mental process regarding a particular annotation category. Annotated text is interpreted automatically for constructing detailed scene descriptions. This includes identifying which scenes to visualize, and identifying the contents and behaviour in each scene. Entity behaviour in a 3D virtual environment is formulated using time-based constraints that are automatically derived from annotations. Constraints are expressed as non-linear symbolic functions that restrict the trajectories of a pair of entities over a continuous interval of time. Solutions to these constraints specify precise behaviour. We create an innovative quantified constraint optimizer for locating sound solutions, which uses interval arithmetic for treating time and space as contiguous quantities. This optimization method uses a technique of constraint relaxation and tightening that allows solution approximations to be located where constraint systems are inconsistent (an ability not previously explored in interval-based quantified constraint solving). 3D virtual environments are populated by automatically selecting geometric models or procedural geometry-creation methods from a library. 3D models are animated according to trajectories derived from constraint solutions. The final animated film is sequenced using a range of modalities including animated 3D graphics, textual subtitles, audio narrations, and foleys. Hierarchical rule-based learning is evaluated over a range of annotation categories. Models are induced for different categories of annotation without modifying the core learning algorithms, and these models are shown to be applicable to different types of books. Models are induced automatically with accuracies ranging between 51.4% and 90.4%, depending on the category. We show that models are refined if further examples are provided, and this supports a boot-strapping process for training the learning mechanism. The task of interpreting annotated fiction text and populating 3D virtual environments is successfully automated using our described techniques. Detailed scene descriptions are created accurately, where between 83% and 96% of the automatically generated descriptions require no manual modification (depending on the type of description). The interval-based quantified constraint optimizer fully automates the behaviour specification process. Sample animated multi-modal 3D films are created using extracts from fiction books that are unrestricted in terms of complexity or subject matter (unlike existing text-to-graphics systems). These examples demonstrate that: behaviour is visualized that corresponds to the descriptions in the original text; appropriate geometry is selected (or created) for visualizing entities in each scene; sequences of scenes are created for a film-like presentation of the story; and that multiple modalities are combined to create a coherent multi-modal representation of the fiction text. This research demonstrates that visual descriptions in fiction text can be automatically identified, and that these descriptions can be converted into corresponding animated virtual environments. Unlike existing text-to-graphics systems, we describe techniques that function over unrestricted natural language text and perform the conversion process without the need for manually constructed repositories of world knowledge. This enables the rapid production of animated 3D virtual environments, allowing the human designer to focus on creative aspects.

Em um ambiente virtualizado, a distribuição de recursos de processamento entre as máquinas virtuais deve ser baseada nos aspectos requeridos pela carga de trabalho executando dentro de cada instância. Entretanto, a imprevisibilidade destas cargas de trabalho, em conjunto com a estratégia de compartilhamento equitativo adotada pelo escalonador no hipervisor, impactam a qualidade de serviço percebida por máquinas virtuais orientadas a eventos de entrada e saída de rede e processamento. Esta dissertação conduz um estudo para estimar a variação da qualidade de serviço percebida por esta classe de máquinas virtuais em cenários de consolidação de servidor. Posteriormente, é proposto um modelo de escalonadores coexistentes, que fundamenta-se no agrupamento de máquinas virtuais com cargas de trabalho similares. A validação deste modelo demonstra que o agrupamento oportuno de máquinas virtuais diminui a variação da qualidade de serviço ocasionada pela concorrência por recursos de processamento, atingindo níveis semelhantes aos cenários sem concorrência. / In a virtualized environment, time slice allocation should be based on virtual machines workloads requirements. However, the unpredictability of these workloads, combined with fair share nature of scheduling strategies, may impact the perceived quality of service of network I/O and CPU intensives virtual machines under server consolidation. This dissertation presents an in-depth analysis of quality of service variation under several scenarios, in order to assess this impact. Furthermore, a coexisting schedulers model, which aims to cluster virtual machines with similar workloads, is proposed as the main contribution of this research. The model validation argues that proper clustering based workload reduces the quality of service variation under server consolidation. In fact, the perceived quality of service under this model resemble a near native scenario.

Les développements récents de la réalité virtuelle font du travail collaboratif assisté par ordinateur un outil prometteur et flexible. Il est en effet aujourd’hui possible, de représenter les données ainsi que les utilisateurs eux-mêmes de manière vivante dans les environnements virtuels collaboratifs (EVC). Les EVC se définissent comme des mondes virtuels distribués, générés par ordinateur, dans lesquels les utilisateurs peuvent se rencontrer, communiquer et interagir entre eux, mais aussi avec des données et des objets 3D. Les utilisateurs peuvent être éloignés physiquement, parler des langues différentes et utiliser des systèmes informatiques hétérogènes tout en collaborant malgré tout au sein d’un EVC. L’objectif principal des EVC est de proposer une collaboration fluide entre plusieurs utilisateurs. Ceci implique de prendre en charge un nombre considérable d’échanges, de communications et de négociations mais également de permettre des activités collaboratives. Par ailleurs, il est nécessaire de proposer des techniques d’interaction ainsi que des moyens pour bien prendre conscience de tout ce qui se passe dans l’environnement. Afin de préserver ces différents aspects dans la conception des EVC, nous nous intéressons à quatre facteurs essentiels : l’immersion, la conscience, la communication et l’intuitivité. Ces facteurs sont déterminants pour le succès des systèmes virtuels collaboratifs. En tenant compte des quatre facteurs cités cidessus, nous proposons et évaluons de nouvelles métaphores pour la navigation et la manipulation afin d’améliorer et d’enrichir les techniques d’interactions dans les EVC. Premièrement, nous proposons et évaluons un ensemble de trois métaphores de navigation pour explorer un environnement à plusieurs : indiquer un chemin en dessinant des flèches, illuminer un chemin à suivre et orienter une boussole pour montrer une direction. Ces métaphores peuvent être implémentées dynamiquement et utilisées directement dans n’importe quels environnements. Nos résultats révèlent que ces métaphores de navigation réduisent considérablement le temps passé à chercher un chemin dans des contextes d’exploration collaborative. Par ailleurs, nous avons développé une technique de manipulation directe dédiée aux environnements virtuels immersifs. Cette technique, qui utilise sept points de contrôle, affranchit l’utilisateur de plusieurs difficultés souvent rencontrées telles que le tremblement de la main ou l’effet Heisenberg lors de la manipulation d’objets 3D dans un système de projection immersive. En réduisant le nombre de degrés de liberté de l’objet manipulé à l’aide de point de contrôle, notre technique permet à l’utilisateur de contrôler partiellement l’objet, rendant ainsi la manipulation d’objets volumineux plus aisée. Enfin, nous avons implémenté et évalué deux métaphores d'interaction dans une application de conception et d'aménagement de poste de travail industriel. En tenant compte des quatre facteurs cités ci-dessus, nous avons implémenté une application de conception de poste de travail pour trois principaux types d'utilisateurs : un utilisateur final, un ingénieur et un expert en ergonomie. Afin d'avoir un niveau d'immersion optimal et des interactions naturelles entre des utilisateurs, nous avons implémenté une configuration de système asymétrique pour chaque utilisateur. Chacun pouvait utiliser quelque métaphores de communication implicites qui étaient simples, naturelles, et pertinentes dans notre contexte collaborative. / Recent progress of the virtual reality technology gives the computer supported collaborative work a potential and flexible tool of vividly representing data as well as users themselves in collaborative virtual environments. Collaborative virtual environments have been defined as computer-based, distributed virtual worlds in which people can meet, communicate and interact with others, with data and 3D objects. People may be geometrically far from each other, speak different languages and use heterogeneous computer systems. The ultimate goal in developing collaborative virtual environments is to support a smooth collaboration between multiple users which involves considerable communication and negotiation, cooperative and collaborative activities, interaction techniques, and awareness process. Considering these aspects into the design of a collaborative virtual environment, we are interested in four main factors, including immersion, awareness, communication and naturalness. These factors greatly determine the success of a collaborative virtual system. From the need of improving and completing interaction techniques in CVEs considering the four preceding factors, in this research we propose and evaluate new metaphors for 3D navigation and manipulation techniques. The first contribution of this research is to propose and evaluate a set of three navigation metaphors in a collaborative exploration context, including drawing directional arrows, lighting up path to follow, and orientating a compass to show a direction. These navigation metaphors can be dynamically implemented and used without constraints in any 3D virtual environments. The empirical result of our experiment revealed that these navigation metaphors considerably reduced wasted time in a wayfinding task of a collaborative exploring scenario. We have developed, in the second part of this research, a direct manipulation technique in immersive virtual environments. This manipulation technique deals with some difficulties the user often encounters such as hand jitter or Heisenberg effects while manipulating 3D objects in immersive projection systems. By dividing the number of degrees of freedom of the manipulated object into each handle of our tool, our technique enables a user to partially control the object, making the manipulation of large objects easier in immersive virtual environments. The last contribution of this research is the implementation and evaluation of two interaction metaphors in a digital mock-up application. Taking into account the four factors including immersion, awareness, communication and naturalness, we have built a workstation design application for three main users: an end-user, a design engineer and an ergonomics expert. In order to have an optimal immersion for the whole application and natural interaction between them, we have implemented an asymmetric system setup at each user's site. Each user could use some implicit communication metaphors which were simple, natural and still relevant in our collaborative context.

The virtualization of IT infrastructure enables the consolidation and pooling of IT resources so that they can be shared over diverse applications to offset the limitation of shrinking resources and growing business needs. Virtualization provides a logical abstraction of physical computing resources and creates computing environments that are not restricted by physical configuration or implementation. Virtualization is very important for cloud computing because the delivery of services is simplified by providing a platform for optimizing complex IT resources in a scalable manner, which makes cloud computing more cost effective. Hypervisor plays an important role in the virtualization of hardware. It is a piece of software that provides a virtualized hardware environment to support running multiple operating systems concurrently using one physical server. Cloud computing has to support multiple operating environments and Hypervisor is the ideal delivery mechanism. The intent of this thesis is to quantitatively and qualitatively compare the performance of VMware ESXi 4.1, Citrix Systems Xen Server 5.6 and Ubuntu 11.04 Server KVM Hypervisors using standard benchmark SPECvirt_sc2010v1.01 formulated by Standard Performance Evaluation Corporation (SPEC) under various workloads simulating real life situations.

Self-healing systems promise operating cost reductions in large-scale computing environments through the automated detection of, and recovery from, faults. However, at present there appears to be little known empirical evidence comparing the different approaches, or demonstrations that such implementations reduce costs. This thesis compares previous and current self-healing approaches before demonstrating a new, unsupervised approach that combines artificial neural networks with performance tests to perform fault identification in an automated fashion, i.e. the correct and accurate determination of which computer features are associated with a given performance test failure. Several key contributions are made in the course of this research including an analysis of the different types of self-healing approaches based on their contextual use, a baseline for future comparisons between self-healing frameworks that use artificial neural networks, and a successful, automated fault identification in cloud infrastructure, and more specifically virtual machines. This approach uses three established machine learning techniques: Naïve Bayes, Baum-Welch, and Contrastive Divergence Learning. The latter demonstrates minimisation of human-interaction beyond previous implementations by producing a list in decreasing order of likelihood of potential root causes (i.e. fault hypotheses) which brings the state of the art one step closer toward fully self-healing systems. This thesis also examines the impact of that different types of faults have on their respective identification. This helps to understand the validity of the data being presented, and how the field is progressing, whilst examining the differences in impact to identification between emulated thread crashes and errant user changes – a contribution believed to be unique to this research. Lastly, future research avenues and conclusions in automated fault identification are described along with lessons learned throughout this endeavor. This includes the progression of artificial neural networks, how learning algorithms are being developed and understood, and possibilities for automatically generating feature locality data.

A virtualização é uma tecnologia chave para a computação em nuvem que permite fornecer recursos computacionais, em forma de máquinas virtuais, para o consumo de serviços de computação. Nos ambientes de computação em nuvem, é importante manter sob controle a alocação de máquinas virtuais nos servidores físicos. Uma alocação adequada implica na redução de custos com hardware, energia e refrigeração, além da melhora da qualidade de serviço. Hipervisores recentes implementam mecanismos para reduzir o consumo de memória RAM através do compartilhamento de páginas idênticas entre máquinas virtuais. Esta dissertação apresenta um novo algoritmo de alocação de máquinas virtuais que busca o equilíbrio no uso dos recursos de CPU, memória, disco e rede e, sobretudo, considera o potencial de compartilhamento de memória entre máquinas virtuais. Através de simulações em cenários distintos, verificou-se que o algoritmo é superior à abordagem padrão na questão do uso equilibrado de recursos e que, considerando o compartilhamento de memória, houve um ganho significativo na disponibilidade deste recurso ao final das alocações. / Virtualization is a key technology for cloud computing, it provides computational resources as virtual machines for consumption of computing services. In cloud computing environments it is important to keep under control the allocation of virtual machines in physical servers. A good allocation brings benefits such as reduction costs in hardware, power, and cooling, also improving the quality of service. Recent hypervisors implement mechanisms to reduce RAM consumption by sharing identical pages between virtual machines. This dissertation presents a new algorithm for virtual machines allocation that seeks the balanced use of CPU, memory, disk, and network. In addition, it considers the potential for sharing memory among virtual machines. Simulations on three distinct scenarios demonstrate that it is superior to the standard approach when considering the balanced use of resources. Considering shared memory, there was an appreciable gain in availability of resources.

Este trabalho apresenta uma arquitetura para simulação de circuitos de lógica com binacional de topologia arbitrária, visando interfaceamento com algoritmos evolutivos para fins de geração de hardware. A implementação é em FPGA utilizando a técnica VRC. O simulador permite circuitos compostos por LUTs de número de entradas parametrizável. A livre interconectividade entre as LUTs permite a construção de circuitos cíclicos. A arquitetura é modular e de interfaceamento simples. Alta performance é obtida através do uso de múltiplos módulos de simulação em paralelo, trazendo resultados que ultrapassam os obtidos em outros trabalhos utilizando DPR. / This work presents an architecture for simulation of combinational logic circuits of arbitrary topology, meant to be interfaced with evolutionary algorithms for hardware generation. It was implemented in FPGA using the VRC technique. The simulator allows for circuits composed of LUTs of parametrizable number of imputs. The free interconectivity between LUTs allows the construction of cyclic circuits. The architecture is modular and of simple interfacing. High performance is obtained by the use of multiple simulation modules in parallel, bringing results that surpass the ones obtained from other works based on DPR.

Este trabalho apresenta uma arquitetura para simulação de circuitos de lógica com binacional de topologia arbitrária, visando interfaceamento com algoritmos evolutivos para fins de geração de hardware. A implementação é em FPGA utilizando a técnica VRC. O simulador permite circuitos compostos por LUTs de número de entradas parametrizável. A livre interconectividade entre as LUTs permite a construção de circuitos cíclicos. A arquitetura é modular e de interfaceamento simples. Alta performance é obtida através do uso de múltiplos módulos de simulação em paralelo, trazendo resultados que ultrapassam os obtidos em outros trabalhos utilizando DPR. / This work presents an architecture for simulation of combinational logic circuits of arbitrary topology, meant to be interfaced with evolutionary algorithms for hardware generation. It was implemented in FPGA using the VRC technique. The simulator allows for circuits composed of LUTs of parametrizable number of imputs. The free interconectivity between LUTs allows the construction of cyclic circuits. The architecture is modular and of simple interfacing. High performance is obtained by the use of multiple simulation modules in parallel, bringing results that surpass the ones obtained from other works based on DPR.

A crescente demanda por ambientes de virtualização de larga escala, como os usados em datacenters e nuvens computacionais, faz com que seja necessário um gerenciamento eficiente dos recursos computacionais utilizados. Um dos recursos mais exigidos nesses ambientes é a memória RAM, que costuma ser o principal fator limitante em relação ao número de máquinas virtuais que podem executar sobre o mesmo host físico. Recentemente, hipervisores trouxeram mecanismos de compartilhamento transparente de memória RAM entre máquinas virtuais, visando diminuir a demanda total de memória no sistema. Esses mecanismos “fundem” páginas idênticas encontradas nas várias máquinas virtuais em um mesmo quadro de memória física, usando uma abordagem copy-on-write, de forma transparente para os sistemas convidados. O objetivo deste estudo é apresentar uma visão geral desses mecanismos e também avaliar seu desempenho e efetividade. São apresentados resultados de experimentos realizados com dois hipervisores populares (VMware e KVM), usando sistemas operacionais convidados distintos (Linux e Windows) e cargas de trabalho diversas (sintéticas e reais). Os resultados obtidos evidenciam diferenças significativas de desempenho entre os hipervisores em função dos sistemas convidados, das cargas de trabalho e do tempo. / The growing demand for large-scale virtualization environments, such as the ones used in cloud computing, has led to a need for efficient management of computing resources. RAM memory is the one of the most required resources in these environments, and is usually the main factor limiting the number of virtual machines that can run on the physical host. Recently, hypervisors have brought mechanisms for transparent memory sharing between virtual machines in order to reduce the total demand for system memory. These mechanisms “merge” similar pages detected in multiple virtual machines into the same physical memory, using a copy-on-write mechanism in a manner that is transparent to the guest systems. The objective of this study is to present an overview of these mechanisms and also evaluate their performance and effectiveness. The results of two popular hypervisors (VMware and KVM) using different guest operating systems (Linux and Windows) and different workloads (synthetic and real) are presented herein. The results show significant performance differences between hypervisors according to the guest system workloads and execution time.