A typical E-commerce system that is deployed on the Internet has multiple layers that include Web users, Web servers, application servers, and a database server. As the system use and user request frequency increase, Web/application servers can be scaled up by replication. A load balancing proxy can be used to route user requests to individual machines that perform the same functionality. <br><br>To address the increasing workload while avoiding replicating the database server, various dynamic caching policies have been proposed to reduce the database workload in E-commerce systems. However, the nature of the changes seen by the database server as a result of dynamic caching remains unknown. A good understanding of this change is fundamental for tuning a database server to get better performance. <br><br> In this study, the TPC-W (a transactional Web E-commerce benchmark) workloads on a database server are characterized under two different dynamic caching mechanisms, which are generalized and implemented as query-result cache and table cache. The characterization focuses on response time, CPU computation, buffer pool references, disk I/O references, and workload classification. <br><br>This thesis combines a variety of analysis techniques: simulation, real time measurement and data mining. The experimental results in this thesis reveal some interesting effects that the dynamic caching has on the database server workload characteristics. The main observations include: (a) dynamic cache can considerably reduce the CPU usage of the database server and the number of database page references when it is heavily loaded; (b) dynamic cache can also reduce the database reference locality, but to a smaller degree than that reported in file servers. The data classification results in this thesis show that with dynamic cache, the database server sees TPC-W profiles more like on-line transaction processing workloads.

A typical E-commerce system that is deployed on the Internet has multiple layers that include Web users, Web servers, application servers, and a database server. As the system use and user request frequency increase, Web/application servers can be scaled up by replication. A load balancing proxy can be used to route user requests to individual machines that perform the same functionality. <br><br>To address the increasing workload while avoiding replicating the database server, various dynamic caching policies have been proposed to reduce the database workload in E-commerce systems. However, the nature of the changes seen by the database server as a result of dynamic caching remains unknown. A good understanding of this change is fundamental for tuning a database server to get better performance. <br><br> In this study, the TPC-W (a transactional Web E-commerce benchmark) workloads on a database server are characterized under two different dynamic caching mechanisms, which are generalized and implemented as query-result cache and table cache. The characterization focuses on response time, CPU computation, buffer pool references, disk I/O references, and workload classification. <br><br>This thesis combines a variety of analysis techniques: simulation, real time measurement and data mining. The experimental results in this thesis reveal some interesting effects that the dynamic caching has on the database server workload characteristics. The main observations include: (a) dynamic cache can considerably reduce the CPU usage of the database server and the number of database page references when it is heavily loaded; (b) dynamic cache can also reduce the database reference locality, but to a smaller degree than that reported in file servers. The data classification results in this thesis show that with dynamic cache, the database server sees TPC-W profiles more like on-line transaction processing workloads.

Submitted by JÚLIO HEBER SILVA (julioheber@yahoo.com.br) on 2017-05-19T17:06:06Z No. of bitstreams: 2 Dissertação - Alex Rabelo Ferreira - 2017.pdf: 1520702 bytes, checksum: 4d6d16d4b1045c459e8d23fc2f6a4c69 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2017-05-22T10:41:24Z (GMT) No. of bitstreams: 2 Dissertação - Alex Rabelo Ferreira - 2017.pdf: 1520702 bytes, checksum: 4d6d16d4b1045c459e8d23fc2f6a4c69 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-05-22T10:41:24Z (GMT). No. of bitstreams: 2 Dissertação - Alex Rabelo Ferreira - 2017.pdf: 1520702 bytes, checksum: 4d6d16d4b1045c459e8d23fc2f6a4c69 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2017-04-25 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / In large-scale data centers, power management techniques such as Dynamic Voltage/Frequency Scaling (DVFS), virtual machine consolidation, and power-capping mechanisms promise impressive energy savings compared to traditional resource management strategies. Most of these techniques rely on coarse-grained monitoring of the workload behavior to apply their optimizations. However, coarse-grained monitoring and black box observations are not satisfactory for predicting the behavior of bursty workloads such as those observed in enterprise, Web servers. In this work, we propose an analytical model to estimate the energy consumption of multi-tier Web Systems. Differently from previous works, our model captures the consumption pattern at the level of fine-grained transactions and for each tier of the system. In addition, our model is based only on CPU utilization and server architectural parameters, which can be easily obtained in today’s production environments. We demonstrate the effectiveness of our model in a real-world experimentation environment, based on the TPC-W benchmark. Results show that our model is able to estimate the energy consumption for a Web system with an average relative error of 6.5% in the worst-case scenario, whereas more complex models of the literature present errors within the same order of magnitude. / Em grandes centros de dados, técnicas de gerenciamento de energia como Dynamic Voltage/Frequency Scaling (DVFS), consolidação de máquinas virtuais e mecanismos de limitação de energia prometem grande economia de energia quando comparados a métodos tradicionais de gerenciamento de recursos. A maioria dessas técnicas utilizam mecanismos caixas-pretas para monitorar o comportamento da carga de trabalho. Contudo, esse tipo de monitoramento não é satisfatório para prever os fenômenos de rajadas, comumente encontrados em serviços e aplicações Web. Neste trabalho, propomos um modelo analí- tico para estimar o consumo de energia de Sistemas Web multi-camadas. Diferentemente de outros trabalhos, nosso modelo captura o padrão de consumo desses sistemas na granularidade de transações e para cada camada do sistema. Além disso, nosso modelo se baseia apenas na utilização de CPU e em parâmetros arquiteturais do servidor, os quais podem ser facilmente obtidos nos ambientes de produção atuais. Demonstramos a efetividade do nosso modelo em um ambiente de experimentação real, baseado no benchmark TPC-W. Os resultados obtidos mostram que nosso modelo é capaz de estimar o consumo de energia para um sistema Web com um erro relativo médio de 6,5% para o pior cenário, enquanto modelos mais complexos da literatura apresentam erros com a mesma ordem de grandeza.