Re: [Ntp] Éric Vyncke's No Objection on draft-ietf-ntp-interleaved-modes-05: (with COMMENT)

[Miroslav Lichvar <[e-mail]>]

On Mon, Jun 28, 2021 at 04:07:51AM -0700, Éric Vyncke via Datatracker wrote:
> Éric Vyncke has entered the following ballot position for
> draft-ietf-ntp-interleaved-modes-05: No Objection

Thanks for the review.

> -- Section 2 --
> "  The server MAY separate the timestamps
>    by IP addresses, but it SHOULD NOT separate them by port numbers,
>    i.e.  clients are allowed to change their source port between
>    requests."
> With draft-ietf-ntp-port-randomization in mind, please change the "clients are
> allowed" into "clients are recommended to"

If I didn't miss an update in the recent changes of the draft, the
port-randomization draft still leaves that per-request or
per-association decision to the implementation with no preference, so
I think they can but are not recommended to change the port between
requests.

> "  Both servers and clients that support the interleaved mode MUST NOT
>    send a packet that has a transmit timestamp equal to the receive
>    timestamp"
> Will it always be possible to comply with the above "MUST NOT" ?

I think it will. If the timestamps happen to be equal, one of the
timestamps can simply be adjusted by 1 before the transmission. That's
~1/4 of nanosecond, well hidden in the typical precision of an NTP
clock.

-- 
Miroslav Lichvar