Re: [quicwg/base-drafts] Clarify PADDING vs PING frames (#838)

[janaiyengar <[e-mail]>]

I really like @huitema's idea of sending a PING (which is a one-byte frame) to elicit an ACK. If we do that though, I don't think we want to leave it to implementations... this should be mandated, since it's congestion control behavior that we want to mandate. 

We don't want implementations to be more aggressive when using PADDING, which means we want to mandate counting them towards congestion control, which means they need to be ackable.

So, as I'm thinking about this, I'm getting convinced that what we really want is to pad with PING frames. Basically, both PING and PADDING are one-byte frames, and we can add as many of them as we want in a packet. We want both to elicit an ack. I don't see any behavioral difference between these two. We can also require that we count PING frames towards congestion control if we don't already.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/issues/838#issuecomment-336556006