Learn to implement microservices across a range of common tech stacks.

Scalability & System Design for Developers

In this module, we’ll look at some fundamental concepts for implementing microservices. Within each concept, learn about the different technologies used to implement it. The technologies include front-end integration with Edge Side Includes (ESI), asynchronous microservices with Kafka and REST feeds, synchronous microservices with the Netflix stack, and Consul, Docker, Kubernetes, Cloud Foundry. Each technology is described and then demonstrated with real code.

Microservice Architecture: Practical Implementation

Kubernetes offers a runtime environment for Docker containers and is very flexible.


# MOMs in Kubernetes

The example in this chapter uses communication with **REST**. Of course, it is also possible to operate a MOM like Kafka ([chapter 7](https://www.educative.io/courses/microservice-architecture-practical-implementation/introduction-B15YQD66Qrk)) in Kubernetes. 

However, MOMs store transmitted messages to guarantee delivery. Kafka even saves the
complete history. Reliable storage of data in a Kubernetes cluster is
feasible, but not easy. Using a MOM other than Kafka does not solve
the problem. 

All MOMs store messages permanently to guarantee delivery. **For reliable communication with a MOM, Kubernetes has to store the data reliably and scalably**.

# Frontend integration with Kubernetes

Kubernetes can be quite easily combined with frontend integration ([chapter 3](https://www.educative.io/courses/microservice-architecture-practical-implementation/introduction-qVw7grDmwy3)), since Kubernetes does not make any assumptions about the UI of the applications.


Client-side frontend integration does not place any demands on the backend. For server-side integration, a cache or web server must be hosted in a Docker container. 

However, these servers do not store any data permanently, so they can easily be operated in Kubernetes.

# Docker Swarm and Docker Compose


Kubernetes offers a very powerful solution and is further developed by many companies in the container area. 

However, Kubernetes
is **also very complex** due to its many features. 

A cluster with Docker
Compose and Docker Swarm can be a **simpler** but **less powerful**
alternative. However, Docker Swarm and Compose also offer
basic features like **service
discovery** and **load balancing**.

# Docker vs. virtualization

As Kubernetes takes over cluster management, it includes features that virtualization solutions also offer. 

This can also lead to **operational concerns**, as reliable cluster operation is a challenge. Another technology in this area is often viewed critically. When deciding against Kubernetes, Docker can still be used without a scheduler.

But then the Kubernetes features for service discovery, load balancing, and routing are missing. They probably have to be implemented by different means.


In this lesson, we'll look at some variations in Kubernetes.

Variations