Re: [quicwg/base-drafts] max_packet_size in 0-RTT (#3447)

[Kazuho Oku <[e-mail]>]

Regarding the original issue, I think I prefer something different than the two options being suggested.

3. Do not require any behavior on the client-side. Allow servers process packets (or datagrams) larger than the advertised limit.

As stated in my previous comment, there is a risk of seeing inferior performance when the client uses a large packet size that was advertised by the previous connection which is greater than the new max_packet_size, and if the server accepts 0-RTT. But the fact is that by sending a packet that is larger than 1200 bytes, the client is already taking the risk that 0-RTT packets might get dropped (by the network).

Also, I do not see why endpoints have to be required to drop packets (or datagrams) that is greater than the advertised value.

I think we can simply say that max_packet_size (or max_datagram_size) is an advisory value, much like max_idle_timeout.

-- 
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/3447#issuecomment-589092455