Re: [webrtc-pc] (How) does SCTP handle CPU-bound congestion on JavaScript thread? (#2086)

> Oh, so we suddenly started encrypting RTP packets with DTLS? 😏

Those SCTP chunks data retransmissions will use bandwidth, that won't be available for audio and video. Note that audio/video may be hardware accelerated and not affected by the js loop at all.

> Do you want to compare spontaneous deadlocks with an explicit exception on a .send call?

I am speaking about receiving data, not sure how the .send call is involved. And I was referring that without a new API we will be in this situation anyway. Maybe the sctp transmission will not be halted, but the browser will die when having reserved several GBs on intermediary buffers. So yes, I prefer the sctp spontaneous deadlock (Note: I am speaking without an API change).



-- 
GitHub Notification of comment by murillo128
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/2086#issuecomment-458211205 using your GitHub account

Received on Monday, 28 January 2019 16:53:44 UTC