Re: [Ntp] Pool and DNS

[Christer Weinigel <[e-mail]>]

On Sat, 2022-10-22 at 22:34 -0700, Hal Murray wrote:
> One day seemed like a reasonable number.
> 
> RFC 8915, section 6, Suggested Format for NTS Cookies, says:
>     Servers should periodically (e.g., once daily) generate a new
>     pair '(I,K)' and immediately switch to using these values for
>     all newly-generated cookies.
> 
> So one day is what I used in NTPsec.  (and my thinking)

Same here.  Netnod's NTS servers rotate their server keys once per day.

> It's easy to save a few old keys.  We should probably agree on some
> lifetime.
> 
> I'd be happy with 5 or 10.  365 seems unreasonable.

Netnod's servers keep track of four keys, the current one which it uses
to generate new cookies and three more keys which it will still be able
to use to decrypt old cookies.

> 
> NTPsec packages the KE server with ntpd so there is no need for that
> protocol. 
>  I don't know of any split implementations.  So as a practical
> matter, it 
> would be hard to put together a pool using that mechanism with
> available code.

Netnod's servers use a split implementation.  The NTS-KE server runs on
a PC and the NTS timestamping is done in a FPGA (inside an Arista 7130
switch).  Actually we have two redundant sets of servers which share
the same set of keys, so if you use the NTS-KE server at
sth1.nts.netnod.se it will generate cookies which work with the NTS
timestapming server found at sth2-ts.nts.netnod.se.

  /Christer