Re: [dnssd] Next steps for privacy discovery

[Christian Huitema <[e-mail]>]

On 10/28/2018 6:48 PM, Lanlan Pan wrote:
>
>
> Christian Huitema <[e-mail]
> <mailto:[e-mail]>>于2018年10月26日周五 下午3:09写道：
>
>     ...
>
>     4) If we kept the current "two phase" structure, use a TLS protocol
>     extension to demonstrate knowledge of the server's public key in the
>     client hello. I think we can build on the work done for SNI
>     encryption,
>     which would fit quite well.
>
>
> I wonder if we could consider about use tls psk (pre-shared key) ? 
> server can assign different key to different client.

That's what we were specifying in the current privacy draft. There are
two issues. The first one is that if you use TLS-PSK, the Client Hello
must include a key identifier. If there is a different key for each
client, the key identifier could become a client identifier. We solved
that  by using a "predictable nonce" for key identifier. The second
issue is of course that assigning different keys to different clients
requires extra management at the server, something that is not needed in
the "private public key" class of solutions.

-- Christian Huitema