Utilizing messaging systems grants us the capability to decouple services from each other: downtime of a service consuming messages does not impact the functionality of the service sending messages and vice-versa. In this talk we will discuss how to setup and use messaging systems. As practical examples, we use AMQP backed by ArtemisMQ, as well as kafka to send and receive messages, automatically as well as programmatically.

1. IBM & Consol present: The day after tomorrow
Messaging with Quarkus
Dr. Marco Bungart
Senior Software Engineer
2021-09-30

2. $> whoami

3. Slide 3
Dr. Marco Bungart
consol.com

2008 – 2013: Studied Bioinformatics/Computer Science
in Jena

2013 – 2018: Ph.D. student in Kassel

Since 2018: Software Engineer at ConSol

Twitter, github, bitbucket, stackoverflow, … : turing85

Interests: Keycloak, GraalVM, Quarkus

4. Quarkus!

5. Slide 5
What is Quarkus?
consol.com

“A Kubernetes Native Java stack tailored for OpenJDK HotSpot and
GraalVM, crafted from the best of breed Java libraries and
standards.” - quarkus.io

Lightweight in execution (small memory footprint)

Compilation to native binary possible through GraalVM

Even lower memory footprint

Single instance may have worse performance, but more
deployments possible due to lower resource consumption
→ net gain

Up & running within minutes
Source:
https://1000logos.net/sonic-logo/

6. Slide 6
Why Quarkus?
consol.com

Developer-joy front & center

Developed by developers, for developers

Excellent Maven- and gradle-integration

Powerful dev-mode, tight feedback loop

Curated list of dependencies, “best of the breed”

Red Hat- and Community-developed

Apps (re)start within milliseconds

7. Messaging

8. Slide 8
Messaging
consol.com

Loosely coupled systems

Do not use when strong coupling is needed/wanted!

Realized through

Microprofile

implementation is asynchronous&reactive, easy to swap between
different providers

Supports JMS/AMQP, Kafka

Vendor-specific implementations (e.g. Qpid, Artemis)

More fine-grained controll, but also more configuration

9. Slide 9
Messaging – What we will look at today
consol.com

Microprofile-base implementation

Messaging-Provider:

ArtemisMQ

Kafka

Communication between two services (customer and order)

Observability

10. Slide 10
Messaging – What we will not look at today
consol.com

Vendor-specific implementations

Kafka-Streams (other usecase)

Partitioning

No demo on a cluster today :( but if you want to use a cluster, try
the IBM Cloud!
https://ibm.biz/BdfGea

11. Architecture of the
system

12. Slide 12
Architecture of the system
consol.com

Customer-Service

Publishes message when a new customer is created (containing
email of customer)

Receives message when a new order for a customer is created

Order-Service

Orders need customers, so customers must be received from
Customer-service before an order can be created

Publishes message when a new order is created (containing email
of customer, an order-id, creation-date and total sum)

Customer-messages via AMQP

Order-messages via Kafka

13. Slide 13
Code Examples
consol.com

Can be found at https://github.com/ConSol/quarkusMessaging

Slides are included

Commits + messages are meant as a step-by-step guide for what
we are going to do

14. Slide 14
Code Examples
consol.com

Can be found at https://github.com/ConSol/quarkus-devmode

Slides are included

Commits + messages are meant as a step-by-step guide for what
we are going to do

15. Slide 15
Live Demo
consol.com
Source: https://makeameme.org/meme/lets-pray-to-9000c9f697

16. Questions?

17. Thank you!

18. ConSol
Consulting & Solutions Software
GmbH
Office Düsseldorf
Kanzlerstraße 8
D-40472 Düsseldorf
Germany
Tel.: +49-89-45841-100
Marco.Bungart@consol.de
www.consol.com
Twitter: @turing85