Real-Time Data Delivery on Spring Boot Using ActiveMQ and STOMP over WebSockets - Part 1
We live in the age of digitized social interaction. As demonstrated by the countless gameplay video streams of games over 20 years old, we feel the need to socialize with even the most static digital content. Now, everything is collaborative, en masse and in real-time, so web apps need to be able to handle this concurrent and user-scoped data in an elegant fashion in order to survive.
The traditional approach to handing out clients data in real-time (or quasi-real-time) was to make clients poll the server at regular intervals. This could be achieved by a simple polling (say, every 5 seconds), or long-polling (same as the other, but with longer-lasting sockets and late responses), but now there are new kids in town. The new approach to this situation is to let the server
push the data using a publish/subscribe mechanism, rather than the client
pulling them in, and technologies like
Server-Sent Events were created for this exact purpose.
For a downloadable version of this code visit the Github repository at https://github.com/ygunayer/realtime-guestbook
First, let me quickly explain what
STOMP is. As you might already know, WebSocket is a pretty low-level transport mechanism, so in order to do RPC over WebSocket you have to roll your own message format, both for the call and for the result. This might be tricky and time-consuming, especially on the serialization part, and
STOMP, which stands for Streaming Text Oriented Messaging Protocol, aims to fix this very problem by defining a few commands. The
spring-messaging package provides the necessary classes to implement this protocol, so we’ll be using that to simplify our work. If you’re interested in more information on STOMP, visit the STOMP website or the relevant Wikipedia page.
Now, let’s model our demo. We’ll want to have a very simple guest book system where users will anonymously enter messages which will then be delivered in real-time to other visitors, but since it’s not the objective of this article, these messages will not be persisted. In order to create such an app, we’ll first have to create a build script (i.e.
build.gradle) with the following dependencies:
Here’s an example for Gradle:
Then, like with any other Spring Boot project, we’ll define an application class to serve as the entry point.
Then, we’ll configure the WebSocket endpoints.
Then, we’ll define a simple controller to accept incoming messages and relays them to listening WebSocket clients using STOMP.
Finally, we’ll create a very simple page to serve as the UI. Notice how the messages are appended when they’re received from the WebSocket listener, rather than after they’ve been sent.
And, that’s it! Here’s a simple demonstration using two different browser windows:
As you can see, these messages are global so everyone can see them. To protect these messages, we’ll be using STOMP’s user-scoped message delivery utilities to target specific users in the next article, so please stay tuned!
Update: The next article is up! Read it at here: Real-Time Data Delivery on Spring Boot Using ActiveMQ and STOMP over WebSockets - Part 2”