/application-modernization-javaee-quarkus

Application Modernization Sample - From Java EE (2010) to Cloud-Native (2021)

Primary LanguageJavaApache License 2.0Apache-2.0

Application Modernization - From Java EE in 2010 to Cloud-Native in 2021

This sample demonstrates how to modernize 10+ year-old Java EE applications with cloud-native technologies and the modern Java runtimes Quarkus and Open Liberty. The modernized application is deployed to and operated with OpenShift, the enterprise distribution of Kubernetes.

Project Structure

Demonstrated Capabilities, Technologies and Tools

Application modernization is a journey. The legacy application has been modernized in ten steps.

  1. Monolith - WebSphere Traditional 8.5.5
    • Java EE 6 app from 2008 with Db2 database running in VMs or bare metal.
  2. Monolith - WebSphere Traditional 9 in container
    • Application converted with Transformation Advisor.
  3. Monolith - WebSphere Liberty
    • Application converted with Eclipse Migration Tools.
  4. Separated frontend
    • Dojo frontend in separate container.
  5. Monolith - Open Liberty
    • Modern project structure.
  6. Strangled Catalog Service and remaining Open Liberty Monolith
    • Strangled catalog service (inspired by Mono2Micro) developed with Quarkus and Postgres.
    • Event driven architecture via Kafka.
  7. Strangled Catalog Service and remaining Quarkus Monolith
    • Strangled Quarkus catalog service uses reactive programming model.
    • Remaining Quarkus monolith runs as native executable.
  8. Micro frontend based web application
    • Developed with single-spa and Vue.js. Messaging is done via RxJS.
  9. CI/CD via Tekton
    • Target OpenShift. Via OpenShift Pipelines operator.
  10. CI/CD via Tekton and ArgoCD
    • Target OpenShift. Via OpenShift GitOps operator.

Introduction Video

The following 12 minutes video describes the project on a high level:

Video

Architecture

Architecture of the legacy application:

Screenshot of legacy storefront application:

Architecture of the modernized application:

Screenshot of modernized storefront application:

Documentation

I've written a series of blogs about this project:

Over the next weeks I've planned to write these blogs:

  • Deployments of Apps to OpenShift via Source
  • Exernalizing Configurations for OpenShift Deployments
  • Deploying Nginx on OpenShift
  • Using CORS for OpenShift Applications
  • Deployments to OpenShift via local Scripts
  • Deploying Tekton on OpenShift
  • Sample Tekton Pipelines for Microservices
  • Debugging Tekton on OpenShift
  • Using Git in Tekton Tasks on OpenShift
  • Deploying ArgoCD on OpenShift
  • Using GitOps on OpenShift

Deployment via Docker Desktop

If you want to run the modernized application locally, you can invoke the following commands. All you need is a local Docker installation and the git CLI.

Notes:

  • Docker requires 14 GB memory, 10 CPUs and 80 GB disk space
  • It takes roughly 15 - 20 minutes to start everything
  • Make sure docker-compose is also installed (sounds like this needs to be installed separately on some systems)
$ git clone https://github.com/nheidloff/application-modernization-javaee-quarkus.git && cd application-modernization-javaee-quarkus
$ ROOT_FOLDER=$(pwd)
$ sh ${ROOT_FOLDER}/scripts-docker/build-and-run.sh

The 'build-and-run.sh' script will launch the following containers.

Once everything has been started, you can open the web applications:

Add the item "Return of the Jedi" to the shopping cart via drag and drop.

Update the price of this item:

$ curl -X PUT "http://localhost/CustomerOrderServicesWeb/jaxrs/Product/1" -H "accept: application/json" -H "Content-Type: application/json" -d "{\"id\":1, \"price\":50}"

Open the "Order History" tab to see the updated price. The new price has been updated in the catalog service and the remaining monolith.

Deployment to OpenShift on IBM Cloud with Tekton

The following scripts deploy the modernized application on Red Hat OpenShift on IBM Cloud. However the same instructions should work for other OpenShift and OCP deployments, for example CodeReady Containers.

First create an IBM Cloud Account. Then create an OpenShift cluster, for example via the IBM Cloud Dashboard. I've tested classic infrastructure, single zone, OpenShift 4.6.17, b3c.8x32 and 3 worker nodes.

Additionally you need to install Tekton. The easiest option is to use the 'OpenShift Pipelines' operator from the OperatorHub view in the OpenShift Console (screenshots). Simply accept all defaults. No local installations are necessary.

$ git clone https://github.com/nheidloff/application-modernization-javaee-quarkus.git && cd application-modernization-javaee-quarkus
$ ROOT_FOLDER=$(pwd)
$ sh ${ROOT_FOLDER}/scripts-openshift-tekton/check-prerequisites.sh
$ oc login ...
$ sh ${ROOT_FOLDER}/scripts-openshift/deploy-db2.sh
$ sh ${ROOT_FOLDER}/scripts-openshift/deploy-kafka.sh
$ sh ${ROOT_FOLDER}/scripts-openshift/deploy-postgres.sh
$ sh ${ROOT_FOLDER}/scripts-openshift-tekton/deploy-application.sh
$ sh ${ROOT_FOLDER}/scripts-openshift-tekton/show-urls.sh