Re: [Ntp] Antw: Re: Comments on draft-ietf-ntp-data-minimization

>>> Harlan Stenn <stenn@nwtime.org> schrieb am 06.11.2018 um 08:58 in
Nachricht
<998a160e-9f09-a262-4195-f094fa1e722e@nwtime.org>:

> On 11/5/18 11:11 PM, Ulrich Windl wrote:
>>>>> Harlan Stenn <stenn@nwtime.org> schrieb am 06.11.2018 um 05:36 in
>> Nachricht
>> <b8e67f89-5420-1f22-c226-c2b57de172c9@nwtime.org>:
>>> I forgot another substantive issue:
>>>
>>> The Poll interval should be correct, and not zero.  This goes, for
>>> example, to "keeping the flock in sync" when a leap smear is being
>>> applied.erval
>>>
>>> Clients should pay attention to the poll value returned by the server
>>> (which is a general point, and not specific to this proposal).
>> 
>> Good point: Where is specified what the poll interval returned by the
server
>> actually means? (I only found: "Poll: 8-bit signed integer representing
the
>> maximum interval between successive messages, in log2 seconds.")
> I'm wondering if your question is a serious question or not.

Actually there's exactly one "p.ppoll" in RFC 5905. It seems to be described
in section 13: "...poll interval is fixed at 2 s; otherwise, the host poll
exponent hpoll is set to the minimum of ppoll from the last packet received
and hpoll from the poll routine, but not less than MINPOLL or greater than
MAXPOLL."

I wasn't aware that a server can "pull down" a clients polling interval.

Is the appendix normative?

> 
> It's described in 5905.  See 9.2. FXMIT, and A.5.1.1. for more information.

Thanks,
Ulrich

> 
> H
> 
>> Regards,
>> Ulrich
>> 
>>>
>>> H
>>>
>>> On 11/5/18 7:45 PM, Harlan Stenn wrote:
>>>> I am generally supportive of
>>>> https://datatracker.ietf.org/doc/html/draft‑ietf‑ntp‑data‑minimization,
>>>> draft 3, and believe 3 substantive changes are needed, and a clarifying
>>>> item should be addressed.
>>>>
>>>> 1) LI should be set correctly, rather than to zero.  A server that does
>>>> leap smearing needs to know if the client is aware of a pending leap
>>>> second.  If a leap‑second‑aware client sends LI=0 to a leap‑smearing
>>>> server the leap‑smearing server will offer leap‑smeared time, which will
>>>> break expectations.
>>>>
>>>> 2) The second paragraph of 4.1 should be deleted, or at least all text
>>>> regarding zeroing the Origin and Receive Timestamps should be removed.
>>>> A conforming client will already have these fields as zero, so this
>>>> additional language is at best confusing or misleading.
>>>>
>>>> 3) When the Transmit Timestamp is set to a random number the client must
>>>> remember this random number and use it for the loopback check.  The
>>>> client must also remember the correct transmit timestamp in order to do
>>>> the math and processing with the response packet.  Please add a section
>>>> that includes this information.
>>>>
>>>> For clarification, please add a definition/explanation of "blind
>>>> spoofing" so people will know exactly what is going on.
>>>>
>>>
>>> ‑‑ 
>>> Harlan Stenn <stenn@nwtime.org>
>>> http://networktimefoundation.org ‑ be a member!
>>>
>>> _______________________________________________
>>> ntp mailing list
>>> ntp@ietf.org 
>>> https://www.ietf.org/mailman/listinfo/ntp 
>> 
>> 
>> 
>> 
> 
> -- 
> Harlan Stenn <stenn@nwtime.org>
> http://networktimefoundation.org - be a member!