CapRtc

CapRtc is a custom-based QoE parameter capture tool for WebRTC traffic.

The tool works by modifying the browser environment. It overrites the native RTCPeerConnection by addition of more event-handler. These event-handlers send the data we require in a periodic manner, which is then available for Download from the background.html page.

Usage

Using the chrome-extension:

Go to chrome://extensions/ and check the box for Developer mode in the top right.
Click on Load Unpacked Extension and select the current folder into it.
Now this extension would be visible on your extensions section in chrome.
Click on the extension. This should open a background page. There is a download button on that page
Now open the webRTC application you wish to collect the stats for.
Click on the Download button on the background page to download all the collected stats as a json.
Before opening a new application. Click on Reset Stats to clear the currently accumulated data.

Collecting stats on the User-Device (Android Chrome Browser):

Setup your device for remote debugging
Once everything is setup and open the console for the session in your android browser.
Paste the code in the file Remote.js directly into remote-console.
Test the webRTC app now. You can also disconnect the User Device. But donot close the web-page before collecting the stats.
To gather the statistics, reconnect your device and the console. Inspect the same webpage on which the RTC App is open.
Type DownloadData() into the console. The data.json file will start downloading on your deive. ( You might get a pop-up on the device for Download based on your permissions.)

Plotting the Data: (deprecated)

Create a New Folder. Copy the plot.py and data.json inside it.

Open a terminal inside the folder and run the script using python.

  python plot.py <Path_To_Data.Json> <Time_Interval (Put 1 here currently)>

This will generate numerours .png files each giving out a graph for the specific parameter.
You may want to elimate graphs which represent no meaningful data.

Now the Default

Plotting webrtc-internals Data

Create a Folder which will store all the generated plots.
Store all the data dumps generated from webrtc-internals page (wont work on files generated from the tool) in a new folder.

Open a terminal inside the folder and run the script using python.

  python scripts/plotInternals.py <Path to Dir Containing data dumps> <Dir to save generated Plots>

Again numerous .png files representing time series graphs will be generated. YOu may want to discard cases which represent no meaningful data.

To plot all the graphs corresponding to a data-dump file use:

  python scripts/plotInternals.py <Path to Data Dump File> <Plot Dir>

This will generate all the plots and save them inside <Plot Dir>/<Name of Data File>/

Creating CSV Data File

To create a csv file :

  python scripts/rawData.py <Path to webrtc dump> <Path to New Data File>

This will generate a csv/ excel file for the data in the format

  param1, param2, ....
  data1, data2 ...
  data1, data2 ...

LICENSE

MIT

This project heavily follows from webRTC-externals which is licensed under MIT.

dak-x/cap-rtc