Building High Availability Infrastructure in Cloud
Mai, Khoa (2017)
Metropolia Ammattikorkeakoulu
2017
Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2017113019084
https://urn.fi/URN:NBN:fi:amk-2017113019084
Tiivistelmä
Software development has always been a race. Companies want to release new products to users as soon as possible. Therefore, nowadays most software development companies implement Continuous Delivery and Infrastructure as Code practices to not only shorten the release process but also to ensure the quality of the software applications.
The main goal of the project was to build a high availability infrastructure in cloud for a microservices backend architecture. The study is based on the principles of Microservices architecture and Continuous Delivery and Infrastructure as Code software engineering practices. The process of creating a deployment pipeline was studied. Two types of deployment, EC2 instances and Containers, were studied to deploy the microservices.
Based on the results, Amazon Web Services was selected as cloud provider. Suitable tools and services were selected to build the deployment pipelines. For most of the use cases, the deployment of Containers was selected, but there were situations where EC2 instances was the most suitable choice.
The main goal of the project was to build a high availability infrastructure in cloud for a microservices backend architecture. The study is based on the principles of Microservices architecture and Continuous Delivery and Infrastructure as Code software engineering practices. The process of creating a deployment pipeline was studied. Two types of deployment, EC2 instances and Containers, were studied to deploy the microservices.
Based on the results, Amazon Web Services was selected as cloud provider. Suitable tools and services were selected to build the deployment pipelines. For most of the use cases, the deployment of Containers was selected, but there were situations where EC2 instances was the most suitable choice.