Decide where the lunchbus goes. Play project to learn Spring Boot + websockets + AngularJS 1.5 + Groovy + Redis.
My goal is to learn Spring Boot and websockets primarily, and also to see what does an application relying solely on websockets look like and its strengths/weaknesses.
- IOU list (done)
- Places list, with tags
- People going
- Chatbox
- Choose a place randomly, restricting by preferences (tags) of those going
- Optionally use Redis (at localhost) as a backing store
Also totally unrelated to the lunchbus functionality, I have in there a log tail demo where a generator generates (at random intervals) about 10 messages a second and sends them in batches 4 times a second to the client, which displays them in a scrolling buffer, as a test case for high-throughput situations with websockets, and to make sure even under load the application still does well. The log generation is only triggered if you click the link to show the logs, otherwise there is no overhead from it.
You need Java 8 JDK installed on your system. While the project is built with Gradle, you do NOT need Gradle installed, as Gradle projects come with a wrapper script that downloads the necessary version of Gradle first if needed.
gradlew bootRun
Then, go to http://localhost:8080/ to see the application.
Another method:
gradlew build
java -jar build\libs\lunchbus-0.1.0.jar
You can also add --server.port=9000 to change the port used to whatever you want.
If you'd like persistence, you can also use Redis. First, you need to install Redis at localhost and run it. It is
highly recommended you don't expose the server to the Internet, so you should enable the "bind 127.0.0.1" option within
the redis.conf file that you use. You can download Redis from http://redis.io/download (or for Windows users, at
https://github.com/MSOpenTech/redis/releases). For example on Windows no installation is needed, just download the zip
file, update redis.windows.conf for the bind setting, then run redis-server redis.windows.conf.
To run lunchbus in Redis mode, just add --app.useRedis=true to the command line. By default it connects to localhost
on the default port 6379. If you want to change this, specify the Redis URL: --app.redisUrl=redis://localhost:6379.
To run in IDE, just run the main class Application. For example in IntelliJ you open the build.gradle file as a project, right click Application and select run.
The repository contains the Procfile and Gradle targets needed to run the application on Heroku. The free Heroku dyno and Heroku Redis addon is sufficient to run the application.
- Websocket interactions are harder to debug than HTTP (REST API) calls, mostly due to fact that browser network panels are so well developed there.
- Websocket-based API doesn't preclude possibility of also having and using a REST API.
- Websocket-based API isn't any harder to develop than REST API, if anything it's a touch easier, and you get instant updates everywhere for free.
- Spring Boot is amazing for rapid prototyping and looks like it can scale well to a full-sized system.
- I still need to test how well this works over the Internet.
- Redis is really fast, it's fast enough that I didn't feel like I needed a local version of data. As a side-effect, the system almost supports clustering out-of-the-box (I would just need to connect Spring boot to the Redis pubsub system to distribute messages across the cluster).
Copyright by Jason Winnebeck, 2016. Licensed under Apache 2.0 License.