/distributed-system-demos

Bachelor Computer Science. Demos for Distributed Systems to try them out.

Primary LanguageJavaMIT LicenseMIT

Distributed System Demos

📂 Java RMI

The Server provides a skeleton with an implemented RMI Interface. This interface can be used from the Client. To use a bidirectional communication, the client provides a stub with the callback interface. With this stub the server can send its answer to the client.

Note: Watch out for the extended classes, because they are the key for RMI.

Class Diagram

Note: Modules can be independent projects where the client and server each have the interface dependency.

RMI class diagram

Usage and Output
  1. Start Server main()
  2. Start Client main()

Server Output:

INFORMATION: Server is started!

Received message: Greetings from Client!
Received message: Greetings from Client!
Received message: Greetings from Client!

Client Output:

Received callback: Client is registered for callback!
Received callback: Server received the message!
Received callback: Server received the message!
Received callback: Server received the message!

System Requirements:

  • Java
  • Maven

📂 MQTT Controller

The controller show a small usage of the Publisher and Subscriber concept of MQTT. The communication goes over a topic on a ActiveMQ Broker.

Class Diagram

Note: Modules can be independent projects.

RMI class diagram

Usage and Output
  1. Run activemq start
  2. Start Subscriber main()
  3. Start Publisher main()

Publisher Output:

Log Date Time server_module.ProducerConnection start
INFO: Started Connection.
Sent message: Hello! at: 1653930520109
Sent message: Hello! at: 1653930521130
Sent message: Hello! at: 1653930522132
Sent message: Hello! at: 1653930523147
Sent message: Hello! at: 1653930524159
Sent message: Hello! at: 1653930525170
Sent message: Hello! at: 1653930526177
Sent message: Hello! at: 1653930527186
Sent message: Hello! at: 1653930528190
Sent message: Hello! at: 1653930529191
Log Date Time server_module.ProducerConnection stop
INFO: Stopped Connection.

Subscriber Output:

Log Date Time client_module.ConsumerConnection start
INFO: Started Connection.
Received: Hello! at: 1653930520109
Received: Hello! at: 1653930521130
Received: Hello! at: 1653930522132
Received: Hello! at: 1653930523147
Received: Hello! at: 1653930524159
Received: Hello! at: 1653930525170
Received: Hello! at: 1653930526177
Received: Hello! at: 1653930527186
Received: Hello! at: 1653930528190
Received: Hello! at: 1653930529191
Received: null
Log Date Time client_module.ConsumerConnection stop
INFO: Stopped Connection.

System Requirements:

  • Java
  • Maven
  • ActiveMQ

📂 Multi Threading

A small example for multi threading is the calculation of Pi in the context of the Monte Carlo problem. New issues like deadlocks and race conditions need to be addressed in manual multi threading.

Usage and Output
  1. Start Application main()

Note: The example was done with compromised processor configuration.

Output:

INFO: Calculate 1.0E8 drops in circle.
INFO: Calculated pi: 3.1414568

Single thread result: 3.1414568
Single thread time: 7.5117718 seconds

INFO: Submitted job: 1
INFO: Submitted job: 2
INFO: Calculate 2.5E7 drops in circle.
INFO: Calculate 2.5E7 drops in circle.
INFO: Submitted job: 3
INFO: Submitted job: 4
INFO: Calculate 2.5E7 drops in circle.
INFO: Calculate 2.5E7 drops in circle.
INFO: Calculated pi: 3.14164704
INFO: Joined result: 1
INFO: Calculated pi: 3.14135488
INFO: Joined result: 2
INFO: Calculated pi: 3.14119488
INFO: Joined result: 3
INFO: Calculated pi: 3.14165968
INFO: Joined result: 4

Multi thread result: 3.14146412
Multi thread time: 36.5720642 seconds

System Requirements:

  • Java
  • Maven

📂 REST API

Monolithic REST API application with Spring Boot.

Table of Content
code
configuration
dependencies
docker compose
Class Diagram

Note: Names might have changed.

Monolithic class diagram

Usage

Development:

  1. Provide the MySQL-Server and the ActiveMQ-Broker on the ports described in the configuration.
  2. Run ShopApplication main()
  3. Open API Documentation on http://localhost:8080/swagger-ui/

Production:

  1. cd ./REST_API
  2. docker-compose up (first start has sometimes race-conditions, with initial exceptions please restart again)
  3. Open API Documentation on http://localhost:8080/swagger-ui/

System Requirements:

  • Java
  • Maven
  • ActiveMQ
  • MySQL
  • (Docker)

📂 REST Microservices

Microservices for an application with Spring Boot. Transformed from the monolithic REST-API. Service Modules are as independent as possible.

Service Code Configuration
Abo Module Abo Code Abo Config
Cart Module Cart Code Cart Config
Payment Module Payment Code Payment Config
Product Module Product Code Product Config
API Gateway Module API Gateway Code API Gateway Config
docker compose
Class Diagram

Note: Diagram of only one service to display the pattern.

Microservice class diagram

Usage
  1. cd ./REST_Microservices
  2. Start Microservice Cluster with docker-compose up (first start has sometimes race-conditions, with initial exceptions please restart again)
  3. Open Service API Documentations

System Requirements:

  • Java
  • Maven
  • ActiveMQ
  • MySQL
  • Docker

📂 REST Client UI

Angular 13-Frontend for all REST APIs. Make sure to set the correct Backend-URL in the environment.

The application represents a shop for products that need frequent payments. The user can add or delete available products, put them into the cart and checkout the cart. The frequent payments will be done in the background and the bank will stack up with the abonnements.

Angular
app root
backend connector
dependencies
dockerfile

Keep in mind there is no real error handling implemented. Keep an eye on the console (Browser F12) if there are some issues.

Usage

Change to the directory: cd ./REST_Client_UI

Development Server:

Run npm install to install the dependencies and ng serve for a dev server. Navigate to http://localhost:4200/. The app will automatically reload if you change any of the source files.

Docker Container:

  1. Run docker build -t client-ui:latest .
  2. Start with docker run -p 4200:4200 --name client-ui client-ui:latest
  3. Open http://localhost:4200/ in the browser.

System Requirements:

  • NodeJS
  • Docker

📑 Kubernetes & Helm

An advanced topic for microservices is the deployment on a Kubernetes Cluster for the container orchestration. To get an idea of Kubernetes, a small demo with MiniKube will be explained.

MiniKube provides a Master Node and Worker Node on the local machine. It requires another virtual machine manager such as Docker, Hyper-V or VirtualBox to do that.

In general the Master Node provides services like an API Server, Scheduler or Controller Manager. The Worker Nodes run the Pods with the Containers and are managed by the master.

For the communication with the API Server of the Cluster, the UI or CLI of Kubernetes can be used as clients. The CLI is provided by Kubectl and already installed with MiniKube.

Components of Kubernetes

Components of Kubernetes

Source: https://kubernetes.io/docs/concepts/overview/components/

MiniKube & Helm Installation

Follow the Official Kubernetes Installation Guide with a comfortable description.

Quick Kubernetes Installation Guide
  1. Install the minikube.exe
  2. Test command: kubectl version
  3. If the command is unknown: (otherwise the installation is finished)
  4. Set System Environment Variables:
  5. Edit > PATH > New: C:\Program Files\Kubernetes\Minikube

Follow the Official Helm Installation Guide with a comfortable description.

Quick Helm Installation Guide
  1. Download the last Helm Release
  2. Unpack it and move the binary to C:\Program Files\Helm
  3. Set System Environment Variables:
  4. Edit > PATH > New: C:\Program Files\Helm
  5. Test command: helm version

System Requirements:

  • Docker

Start MiniKube Cluster

Run following commands to start the cluster:

  1. minikube start
  2. minikube status
  3. minikube dashboard

Additional commands:

  • Start in debug mode: minikube start --alsologtostderr
  • Stop the cluster: minikube delete
  • Basic Controls

Deployment with Helm

The deployment could be done with the Kubectl CLI to access the Kubernetes API Server. This produces a lot of effort for multiple services and applications. To make things easier, Helm can be used.

Helm is a tool that provides a Manager for Helm Charts that bundle YAML Files with different configurations. In combination with the Templating Feature, blueprints can be created to make configurations easier. The values from the Helm Charts will get injected into the final deployment configuration for Kubernetes.

In a nutshell each YAML Files for a Pod gets bundled into a Helm Chart to simplify the configuration and to make it reusable.

Helm 2 vs. 3

In Helm 2 a Tiller Service was implemented in the Kubernetes Cluster. It enabled Release Management to upgrade or rollback in between different versions of the deployment. Because this service had too much power, it became a security issue. Therefore, Helm 3 removed Tiller for more security with the loss of the release management.

Helm Chart Structure

Create new chart: helm create MyChart

MyChart/  
   Chart.yaml     --> meta information
   values.yaml    --> values for the template files
   charts/        --> chart dependencies
   templates/     --> actual template files
   configs/       --> config maps for applications

To deploy the application on the local MiniKube, the containers need to be build fist with the docker-image-build.sh. Then the Helm Chart can be packaged:

The Helm Chart will be available and explained in the future

helm package HelmChart

After that we can install the archive file on the running Kubernetes Cluster:

helm install release1 HelmChart-0.1.0.tgz

💾 System Requirements

Java 11

  1. Download: Java 11+
  2. Install the executable
  3. Set System Environment Variables:
  4. New: JAVA_HOME = C:\Program Files\Java\jdk-11
  5. Edit > PATH > New: %JAVA_HOME%\bin
  6. Test command: java -version

Maven

  1. Download: Maven 3.6.3+
  2. Unzip it to: C:\Program Files\maven
  3. Set System Environment Variables:
  4. New: MAVEN_HOME = C:\Program Files\maven
  5. Edit > PATH > New: %MAVEN_HOME%\bin
  6. Test: $ mvn -v

ActiveMQ

  1. Download: ActiveMQ 5.17+
  2. Unzip it to: C:\Program Files\activemq
  3. Set System Environment Variables:
  4. Edit > PATH > New: C:\Program Files\activemq\bin
  5. Test command: activemq start
  6. Admin page: http://localhost:8161/admin with admin, admin credentials.

MySQL

  1. Download: MySQL 8 Server
  2. Install the executable
  3. Select Server only
  4. Default root password: password (not for productive DB!)

NodeJS

  1. Download: NodeJS
  2. Install the executable
  3. Install the Angular CLI: npm install -g @angular/cli@13.3.9

Docker

  1. Download: Docker Desktop
  2. Install the executable