Voice Links Installation Guide

Caveat: This application is a proof of concept and does not conform to web application security best practices. Caution should be taken before running any of the commands listed in the installation guide. Note that some commands use example.com as a placeholder domain.

This guide has been designed for Ubuntu 14.04.3

Installing EasyRTC

1. Install Node.js with sudo apt-get install nodejs
2. Download EasyRTC v1.0.16 (later versions may work) with git clone -b beta https://github.com/priologic/easyrtc.git
3. Install npm with sudo apt-get install npm
4. cd easyrtc to open the cloned contents and npm install
5. cd server_example to open the server directory folder and npm install
6. Install screen via package manager with sudo apt-get install screen
7. Start a screen session with screen -R easyrtc
8. Run nodejs server.js from inside the server_example directory

Note: You can exit the screen session using ctrl-a + d

You will now have a working installation of EasyRTC running on port 8080. Test this by browsing to it in a supported web browser.

Configuring HTTPS on EasyRTC

Supported browsers require HTTPS when using WebRTC.

1. Create an SSL key and cert, replacing example.com with your domain

   sudo git clone https://github.com/letsencrypt/letsencrypt /opt/letsencrypt
   cd /opt/letsencrypt
   sudo -H ./letsencrypt-auto certonly --standalone -d example.com
   

2. cd ~/easyrtc/server_example and nano server.js

3. Remove the following code

   var http    = require("http");

4. Insert the following code in its place

   var https = require("https");
   var fs = require("fs");

5. Remove the following code

   var webServer = http.createServer(httpApp).listen(8080);

6. Insert the following code in its place, replacing example.com with your domain

   var webServer = https.createServer(
   {
   	key:	fs.readFileSync("/etc/letsencrypt/live/example.com/privkey.pem"),
   	cert:	fs.readFileSync("/etc/letsencrypt/live/example.com/cert.pem"),
   	ca:		fs.readFileSync("/etc/letsencrypt/live/example.com/chain.pem")
   },
   app).listen(8080);

7. Run npm install from inside the server_example directory again

You can now run screen -r easyrtc and sudo nodejs server.js to run EasyRTC with HTTPS support.

Note: You can exit the screen session using ctrl-a + d

Test your configuration works inside a supported browser.

Apache/PHP/MySQL (LAMP) Install

1. Install the required dependancies for Voice Links sudo apt-get -y install apache2 mysql-server php5-mysql php5 libapache2-mod-php5
2. Enter password aREALLYsecurepa$$w0rd for MySQL root password
   • If you change the password make sure to edit config.php with the updated MySQL credentials
3. Run sudo mysql_install_db
4. Run sudo mysql_secure_installation and follow the instructions - defaults work fine

You will now have a standard LAMP stack installed.

Cloning the Web App

Clone the voice-links web application.

1. git clone https://github.com/FireGrey/voice-links.git
2. sudo mv ./voice-links /var/www/voice-links/

Configuring the Web Server

1. Run mysql -u root -p < /var/www/voice-links/DB.sql
2. nano /etc/apache2/apache2.conf and add the following

   <Directory /var/www/voice-links>
   	Options Indexes FollowSymLinks
   	AllowOverride All
   	Require all granted
   </Directory>
   

3. sudo a2enmod rewrite
4. sudo service apache2 restart
5. nano /etc/apache2/sites-available/000-default.conf and add the following, replacing example.com with your domain

<VirtualHost *:80>
	DocumentRoot /var/www/voice-links
	ServerName example.com
</VirtualHost>

6. sudo service apache2 restart
7. cd /opt/letsencrypt
8. Run ./letsencrypt-auto --apache -d example.com replacing example.com with your domain

This will configure Voice Links to run at the base URL example.com/. If you chose to run Voice Links under a different base URL i.e. example.com/some-directory/ you will need to update the RewriteBase in the .htaccess file to match.

Notes

Saved Recordings Filename

Recordings are saved for each user in the room. The filename format is the user ID followed by the offset between the user clicking the Start recording button and the remote peers microphone being recorded. This allows you to accurately merge the individual recorded files even when peers join the room later than when the recording begun.

ID-DTcqZTWJCTPy6EqO_Offset-0ms filename represents the user ID DTcqZTWJCTPy6EqO who was already in the room at the time recording begun because their offset is 0ms zero milliseconds.

ID-OLoodr8X5G1mCvKe_Offset-18927ms filename represents user ID OLoodr8X5G1mCvKe who joined the room 18927ms 18.9 seconds after recording begun.

Browser Support

Chrome 50.0 -- Well Supported

Firefox 46.0.1 -- Moderataly Supported (see bugs)

Bugs

1. No Viable Ice errors for heavily NAT'ed peers
   • Workaround: Switch to a quieter connection (3/4G tether from a mobile phone often works, VPN sometimes works)
2. In Firefox, peers who join after recording has started are not included in recording
   • Waiting for newer version of MediaStreamRecorder (possibly v1.3.3)
3. Rare crash in Firefox when recording multiple peers
   • Hard to replicate
4. Chrome recordings are only saved as webm (inconvient format for audio editing)
   • Waiting for newer version of MediaStreamRecorder (possibly v1.3.3)
5. HTML5 audio box timer freezes in Chrome with more than 2 peers
   • Likely fixed in later versions of Chrome
6. Output selection boxes blank for some Chrome users
   • Workaround: Open chrome://flags and enable Experimental Web Platform
7. Output selection only supported by Chrome
   • Likely supported in later versions of Firefox