Skip to content

Latest commit

 

History

History
49 lines (36 loc) · 2.3 KB

video_stream.md

File metadata and controls

49 lines (36 loc) · 2.3 KB
Raw
layout title
post
Setup the Video Stream

In this part of the instructions you will setup the video stream. First, make sure you have Multipath-TCP (MPTCP) version 0.94.0 installed on your system. We used Ubuntu 16.04 with kernel version 4.14.24. See the official Multipath-TCP website for installation instruction. After that, make sure you set the MPTCP options correctly:

sudo sysctl -w net.mptcp.mptcp_enabled=1
sudo sysctl -w net.mptcp.mptcp_path_manage=fullmesh
sudo sysctl -w net.mptcp.mptcp_scheduler=redundant
sudo sysctl -w net.mptcp.mptcp_checksum=1
sudo sysctl -w net.mptcp.mptcp_syn_retries=3

Up next, install required software, which is docker-ce and docker-compose. For instructions see the guides on their homepage for docker-ce and docker-compose.

After that, download and prepare the video. We provide a shell script for that. Checkout out the following repo and run the download script

git clone https://github.com/umr-ds/seamcon-videostream.git
cd seamcon-videostream
./download_video.sh

This will create a directory ./video/, where all required files are stored.

Our setup reports various statistics about the video stream. Therefore, we deploy a JSON server instance, where these stats are stored in a file called db.json. We prepared an empty example, which should be used. Rename the example file and set a symlink so that JSON server uses the right file.

cp db.json_empty db.json
ln -s db.json .~db.json

Additionally, you have to set the endpoint of your JSON server machine. This is done in the file dashloader/app.js, line 36150. Just replace the API_URL with the IP or the hostname of the machine and the port where the JSON server docker container (which will be started in a second) is running.

Finally, for starting the stream, start all containers with our provided docker compose file.

docker-compose up -d

You should see the video at http://<IP_OF_YOUR_MACHINE>:8080. The stats of the video stream are in the db.json file. Alternatively, you can see them at http://<IP_OF_YOUR_MACHINE>:3000.

Cool, the video stream is running. Now head over to the final step.