Analysis and Implementation of Microservices Using Docker

Implementation of micro-services is among the most challenging and important tasks in the field of Computer Science. For implementing the microservice architecture, Docker has been used in the current scenario. Creating a machine learning model that works on one computer is not really difficult. But creating a model that can scale and run on all types of servers around the world, it's more challenging. This work proposed a way to deploy a certain type of trained model in microservice architecture using Docker to use anywhere. For practical implementation, two models for machine learning has been used. One that addresses the handwritten digit recognition problem which has been an open problem in pattern classification since a long time, and another model resolves the issue of house pricing based on several attributes. The purpose of implementation of machine learning model in microservice architecture using Docker is to enable a method from which anyone can use a machine learning model without worrying for their machine configuration and dependencies of the machine learning model.