Comparison of Deployment Options for Microservices in the Cloud

In the context of digital transformation, organizations must change their systems and technology platforms. Cloud Computing is a powerful technology available at disposal of organizations to accelerate transformation plans. The cloud offers several benefits such as greater agility in development, quick and easy access to information from anywhere and anytime, as well as great savings since you pay for what you use, thus reducing infrastructure investment costs. Another great advantage is that cloud providers offer many options to migrate services and applications. In this sense, it is important to characterize the resource and pattern usage of applications to choose the cloud service that offers the best relation cost/benefit. In this article, we focus on microservices and analyze three options for deploying them in the Google Cloud Platform: virtual machines (Compute Engine), Google Kubernetes Engine and Cloud Functions. Once the services are deployed, we generate artificial load to measure the response times and the number of requirements served. We also calculate the cost of the three options. The results agree with other studies that recommend serverless options such as Cloud Functions over container-based options or virtual machines. While these studies focus primarily on performance and cost, our work also addresses elements related to ease of deployment.