Re: [Ntp] Should NTPv5 have QUIC bindings?

[Tony Finch <[e-mail]>]

Hal Murray <[e-mail]> wrote:
>
> I'm assuming the idea is to replace NTS with QUIC.
>
> One of the goals of NTS was to avoid per-connection storage on the server.  It
> uses TLS to get cookies that hold the connection info.  I don't see how to do
> that with QUIC.

I guess the client could use 0-RTT session resumption, which uses a
session ticket held by the client to retain state (such as the symmetric
crypto key) between queries. I don't know if QUIC allows the server to
close the session immediately with its first response packet, so it can
reply and immediately forget about the client, or if the server will have
to keep state for at least 1 RTT per query.

> I saw mention of QUIC not using the IP Address for a connection key.  That
> sounds like an invitation for easy tracking.  If we take non-tracking as a
> serious goal, that could be enough to knock QUIC out of consideration.

Hmm, I can't find anything useful about the privacy considerations of
0-RTT in QUIC or in TLS/1.3. One advantage of using a popular security
layer ought to be that it deals with issues like this so NTP would not
have to worry at all...

Tony.
-- 
f.anthony.n.finch  <[e-mail]>  https://dotat.at/
defend the right to speak, write, worship, associate, and vote freely