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

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.

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

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!