SCALABLE MICROSERVICES AND CONTAINSERS USING OS-LEVEL VIRTUALIZATION

Abstract

This thesis discusses the microservice architecture and how it promotes the scalability of Web services through containerization. The microservice architecture style is a method of developing a single application into a series of small services, each of which runs in its own process and communicates through standard APIs. The Docker architecture is extensible and can work with other software development tools to implement an extensible construction and deployment system for microservices. This thesis uses Docker Engine, Docker Machine, Docker Compose and Docker Swarm to achieve the scalability of the microservice architecture. Docker Swarm is a middleware in the orchestration layer, which abstracts the complexity caused by the microservice architecture. This abstraction simplifies the design and implementation of microservices, and improves system scalability by eliminating contention lag and minimizing consistency lag. From the results obtained by measuring the container startup delay, it can be seen that Docker swarm scales linearly with the increase in the number of containers. The design of Docker Swarm orchestration software was also found to be based on the microservice architecture, so its linear scalability. The most important factor affecting the scalability of Docker Swarm is the Raft consensus algorithm. This algorithm also plays an important role in distributed databases. In order to check the scalability of the back-end service, we used the Crate database, which runs on a cluster of machines managed by Docker Swarm. It has been observed that Docker Swarm does simplify the scalability of many online web services. Docker swarm-based orchestration is and will continue to be a viable candidate to simplify the scalability of microservices and web services.