Re: [Ntp] NAT devices not translating privileged ports

[Fernando Gont <[e-mail]>]

Hi, Miro,

On Thu, 2021-06-10 at 16:44 +0200, Miroslav Lichvar wrote:
> On Thu, Jun 10, 2021 at 09:37:43AM +0000, Fernando Gont wrote:
> > I'm now considering whether we'd be better off removing the whole
> > Section 3.4.? i.e., remove this:
> 
> That would work for me.
> 
> There is an effect of NAT that could be mentioned if you are
> considering some replacement for the text. NATs typically have a
> shorter timeout for UDP sessions that the client polling interval, so
> if there are multiple clients using the same server behind NAT, their
> source port from the server's point of view will be randomized even
> if
> their local port is the same.


Are you assuming that if there's a "collision" at the NAT, the NAT will
change the service port to some other port?


In any case,

1) noting this would b tricky: it has been noted that in some of those
cases, the NAT may randomize the port, but from the same pool
(privileged vs non-privileged ports).

2) Since the behavior probably varies from one implementation to
another, and since I'm not sure it would be easy to reference such
behavior, I wonder if we'd be better off simply removing the text....

Thanks!

Regards,
-- 
Fernando Gont
Director of Information Security
EdgeUno, Inc.
PGP Fingerprint: DFBD 63E3 B248 AE79 C598 AF23 EBAE DA03 0644 1531