Skip to content

Latest commit

 

History

History
73 lines (51 loc) · 3.27 KB

README.md

File metadata and controls

73 lines (51 loc) · 3.27 KB
Raw

STUNner demo: Magic mirror with Kurento

This is the Kurento Magic mirror demo, adopted for STUNner and Kubernetes. The demo shows a basic WebRTC loopback server with some media processing added: the application uses computer vision and augmented reality techniques to add a funny hat on top of faces.

The demo contains a fully featured WebRTC application server, developed using the express framework for Node.js, and the standard Kurento media server for processing audio/video media. In the original demo, both the application server and the media server are expected to be reachable via a public routable IP address. This demo was adopted here for using with STUNner, so that it can be installed into a standard Kubernetes cluster.

In this demo you will learn the following steps to:

  • integrate a typical WebRTC application server to be used with STUNner,
  • deploy the modified application server into a Kubernetes,
  • deploy the Kurento media server into Kubernetes behind STUNner,
  • scale a standard WebRTC workload using Kubernetes and STUNner.

Installation

Prerequisites

The tutorial assumes a fresh STUNner installation; see the STUNner installation and configuration guide. Create a namespace called stunner if there is none. You need a WebRTC-compatible browser to run this tutorial. Basically any modern browser will do; we usually test our WebRTC applications with Firefox.

Install the Application

Install the WebRTC application and Kurento media servers, altogether with the corresponding services and Kubernetes objects (see the content of the yaml for details):

$ git clone https://github.com/l7mp/stunner
$ cd stunner
$ kubectl apply -f docs/examples/kurento-magic-mirror/kurento-magic-mirror-server.yaml

STUNner configuration

Next, we configure STUNner to act as a TURN server towards clients, and to let media reach the media server.

$ kubectl apply -f docs/examples/kurento-magic-mirror/kurento-magic-mirror-stunner.yaml

The demo exposes a publicly available HTTPS web service on port 8443. Kubernetes assigns an ephemeral public IP address to the web service, so first we need to learn the external IP.

$ kubectl get service webrtc-server -n default -o jsonpath='{.status.loadBalancer.ingress[0].ip}'

The result should be a valid IP address in the form A.B.C.D. If no IP address is returned, wait a bit more until Kubernetes successfully assigns the external IP. Then, direct your browser to the URL https://<A.B.C.D>:8443 (of course, make sure substitute the previous IP address), accept the self-signed certificate, and hit the Start button.

Scaling

This demo uses the AI/ML based computer vision features built into Kurento to process media. As such, it is fairly hard on the media server CPU. Thanks to STUNner, the media server pool can be simply (auto-)scaled with Kubernetes effortlessly.

The below command will scale the Kurento media server pool to 4 instances and again, automatic health-checks and load-balancing should just work as expected.

$ kubectl scale deployment kms --replicas=4

Help

STUNner development is coordinated in Discord, feel free to join.