Informações Principais
     Resumo
     Abstract
     Introdução
     Conclusão
     Download
  
  
  
 
Abstract
 
 
Acadêmico(a): Marco Aurélio Alano Godinho
Título: Arquitetura na nuvem pública AWS voltada à hospedagem de uma aplicação Magento
 
Abstract:
The challenges encountered in the operations industry are often related to the inconsistencies
that arise during the management of a clusters and the entire infrastructure. When there are
multiple teams working on the same project, there are high chances of similar requirements
ending up implemented in different ways. Such situations can impact negatively on the
infrastructure and applications supported, which is especially harmful to the business when
the supported platform directly tied to its financial income. This is the case when it comes to
an e-commerce platform like Magento. This work presents a solution that uses Infrastructure
as Code concepts to create an architecture in the Amazon Web Services (AWS) cloud to
ensure the consistency of the cloud environment, allowing the infrastructure to be addressed
with software engineering techniques. Code evaluation, revision and testing can be applied
before making changes to the production environment. In terms of architecture, the
infrastructure created is tolerant to hardware failure because it supports the loss of an AWS
availability zone and is able to scale-out the computational resources automatically to adapt to
demand. Docker containers were used to run fundamental operations services, such as the
Magento application itself and the Varnish cache layer, which run in an AWS Elastic
Container Service cluster. Other technologies such as relational database, in-memory
database, load balancing, and network resources come from AWS-specific services for these
purposes. The resiliency and scalability goals were achieved and tested simulating the loss of
an availability zone manually and load testing the application with Locust. The infrastructure
created is capable of supporting the creation of 415 quotes per minute in the store.