Selective Performance Optimization of Business Features using Microservices approach

Abstract

Any business application is composed of multiple pieces of functionality. Business requires to know which of those generate most profit to an organization and expects them to perform well and handle increasing load in future. This thesis explores Microservices as an approach that allows to separate valuable business functionality into an independent service, so that problematic parts of a system can be scaled out individually as well as replaced with different implementations in attempts to eliminate or reduce identified performance bottlenecks. Analysis of Microservices architecture, supported by DDD, BDD and Hexagonal architecture, is provided, with emphasis on replaceability and testability that can be used to implement and measure experiments that address performance bottlenecks. Based on the experience gathered from the prototype implementation the identified bottlenecks are discussed as well as implemented solutions to reduce them. Eventually, conclusions derived from chosen optimization approaches and tools are presented.