The increasingly popular cloud-computing paradigm provides on-demand access to computing and storage with the appearance of unlimited resources. Users are given access to a variety of data and software utilities to manage their work. Users rent virtual resources and pay for only what they use. In spite of the many benefits that cloud computing promises, the lack of dependability in shared virtualized infrastructures is a major obstacle for its wider adoption, especially for mission-critical applications. Virtualization and multi-tenancy increase system complexity and dynamicity. They introduce new sources of failure degrading the dependability of cloud computing systems. To assure cloud dependability, in my dissertation research, I develop autonomic failure identification and diagnosis techniques that are crucial for understanding emergent, cloud-wide phenomena and self-managing resource burdens for cloud availability and productivity enhancement. We study the runtime cloud performance data collected from a cloud test-bed and by using traces from production cloud systems. We define cloud signatures including those metrics that are most relevant to failure instances. We exploit profiled cloud performance data in both time and frequency domain to identify anomalous cloud behaviors and leverage cloud metric subspace analysis to automate the diagnosis of observed failures. We implement a prototype of the anomaly identification system and conduct the experiments in an on-campus cloud computing test-bed and by using the Google datacenter traces. Our experimental results show that our proposed anomaly detection mechanism can achieve 93% detection sensitivity while keeping the false positive rate as low as 6.1% and outperform other tested anomaly detection schemes. In addition, the anomaly detector adapts itself by recursively learning from these newly verified detection results to refine future detection.

Orientador: Katia Lucchesi Cavalca Dedini, Hélio Fiori de Castro / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-16T15:47:14Z (GMT). No. of bitstreams: 1 Camargo_LucasWardFrancode_M.pdf: 3927704 bytes, checksum: fae6a20117d598883d434bc95dfb93fd (MD5) Previous issue date: 2010 / Resumo: O estudo de máquinas rotativas ocupa uma posição destacada no contexto de máquinas e estruturas devido à quantidade significativa de fenômenos que podem ocorrer durante a operação desses equipamentos. Este trabalho é focado no estudo de sistemas rotativos sustentados por mancais hidrodinâmicos (com ou sem um acoplamento conectando o eixo movido ao eixo motriz) e, principalmente, na análise das falhas que podem mais comumente ocorrer nesses sistemas (desbalanceamento, empenamento e desalinhamento). A análise é estendida para o caso de múltiplas falhas, visto que essa é uma condição comum em sistemas rotativos. O alinhamento e o balanceamento perfeitos da máquina não podem ser obtidos em aplicações reais e, portanto, forças de excitação são geradas nesses sistemas, as quais afetam significantemente a operação das máquinas. Dessa forma, a influência do desbalanceamento, empenamento e desalinhamento nas amplitudes de vibração constituem uma importante consideração, especialmente o efeito dos componentes harmônicos que podem surgir em uma máquina desalinhada. Para representar matematicamente o sistema, um modelo de elementos finitos é utilizado e a análise é realizada no domínio da frequência. O acoplamento também é modelado como um elemento finito e as forças de desalinhamento, desbalanceamento e empenamento são incluídas no modelo de falhas. Com a adição de um nível de ruído aos sinais simulados, um possível sinal experimental é obtido e, com o objetivo de identificar os parâmetros de falha considerados desconhecidos, um algoritmo genético multiobjetivo é proposto. A análise individual de cada função objetivo (erro entre sinal simulado e sinal ajustado) permite que seja possível identificar um conjunto de soluções ótimas para o problema e, assim, identificar os parâmetros de falhas / Abstract: The study of rotating machinery occupies an outstanding position in the study of machinery and structures due to the significant amount of phenomena that occur in the operation of these equipments. This work is focused on the study of rotating systems supported by hydrodynamic bearings (with or without a flexible coupling connecting the driving motor and the driven shaft), mainly considering faults that can commonly occur in these systems (misalignment, bow and unbalance) as the presence of multiple faults in a real machine is a common situation. The perfect balance and alignment of the machine cannot be achieved in practical applications. Consequently excitation forces are generated at these systems, which significantly affect the operation of the machines. In this way, the influence of the unbalance, bow and misalignment on the vibration amplitude is an important consideration, specially the effect of the harmonic components that can rise when a machine is misaligned. To mathematically represent the system, a finite element model is used and the analysis is held in the frequency domain. The coupling is also modeled as a finite element and the misalignment, bow and unbalance forces are included in the fault model. Noise is added to the simulated results to build a possible experimental result and, in order to identify the fault unknown parameters, the multi-objective genetic algorithm is proposed. The individual analysis of each objective function (difference between simulated and adjusted results) allows the identification of an optimal set of solutions resulting in the identification the fault parameters / Mestrado / Mecanica dos Sólidos e Projeto Mecanico / Mestre em Engenharia Mecânica