Re: [webrtc-pc] Add hardwareAccelerated field to RTCRtpCodecCapability (#2356)

> @aboba at first read, this would seem to be a way for sites to pierce the isolation between the browser, and device hardware, and so adds new fingerprinting surface. Is this information (implicitly or otherwise) gated behind a permission somewhere earlier in the flow that I'm missing?

I'm not sure if I see how this adds something that is not already implicitly known.

**In WebRTC**
In the Chromium implementation of WebRTC there are three included SW decoders - H264, VP8, and VP9. In addition to these codecs, there might be support for some additional codecs/profiles through external HW decoders. So already with the existing specification (without this field), any other codec than the three SW codecs would imply that this is a codec with only HW decoder support. The additional information that is added by this field would be if there's HW support for the three native SW decoders.

However, as soon as the stream starts, the implementation name would reveal this anyway. 

**In Chrome**
If we take a look outside WebRTC, there are at least two other ways of retrieving similar or even more detailed information. 1) Through the MediaCapabilities API which returns supported, smooth, and powerEfficient for any codec configuration that is requested. 2) Through WebGL which provides "Unmasked renderer" which is essentially the name of the graphics card chip (see https://webglreport.com/).


-- 
GitHub Notification of comment by drkron
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/pull/2356#issuecomment-552409803 using your GitHub account

Received on Monday, 11 November 2019 11:40:24 UTC