Re: [Ntp] Antw: Re: Calls for Adoption -- NTP Extension Field drafts -- Four separate drafts

[Harlan Stenn <[e-mail]>]

On 9/2/2019 5:43 AM, Miroslav Lichvar wrote:
> On Mon, Sep 02, 2019 at 01:32:11PM +0200, Magnus Danielson wrote:
>>>> If I only had time to describe all the things I would do to enhance NTP,
>>>> it has a number of built-in assumptions which is not useful or correct.
>>> How about a list of at least some of the built-in assumptions that you
>>> think are not useful or correct?
>>
>> From the top of my head a few:
>>
>> The Allan interception point is based on incorrect theory of the noise.
>>
>> Rather than drive things to highest stratum, going to closes
>> intermediate node and then let that do aggregated questioning to
>> increase the width of single stream rather than separate allows better
>> noise surpression of asymmetric delays.
>>
>> The frequency/phase lock-in methods is dated and has failures in
>> error-handling, especially with how calibration works and can become
>> incorrect. There is simpler methods with much better error-handling.
> 
> If I understand it correctly, those are not a problem of the NTPv4
> format or protocol, but rather RFC 5905. I'm not sure if a new NTP
> version is needed to fix that. A draft could specify that NTPv4
> implementations are free to use different algorithms and models of the
> clock than what is specified in RFC 5905. That would just describe the
> current state. I think it was an idea for the NTPv5 specifications
> that it would have no algorithms, only the message format and
> protocol.

This was discussed for NTPv4 as well.

The thing is that NTP is a complete specification with a goal of time
synchronization that includes *known behavior in the face of impulse
response*.

If one uses different algorithms with different impulse response
behaviors, this can trigger oscillations and even unstable tracking (the
wheels fly off the bus).

Be careful what you wish for.

> What I wanted to collect were issues that may require changes in the
> format itself. Like insufficient resolution of the root delay and
> dispersion, missing timescale identification, minimum length of
> extension fields, etc.
> 

-- 
Harlan Stenn, Network Time Foundation
http://nwtime.org - be a Member!