Performance Evaluation and Comparison of Microservices and Serverless Deployments in Cloud

Abstract

Microservices and serverless are arguably the two most widely used architectures today for deploying applications in the cloud. With both these technologies, applications can take advantage of faster delivery, lightweight, scalable, and lower development and maintenance costs. However, there are ongoing debates concerning which of these two architectures to use for deploying a given application. This paper evaluates and compares them quantitatively in terms of their performance and cost, based on a use case study for an image processing application. The study was conducted by deploying the application using the two technologies in the two major cloud platforms, Amazon AWS and Google Cloud. Results showed that serverless perform better in terms of performance and cost, whereas microservices show superiority in terms of memory use.