[ntp:hackers] refclock servers and the stratum option

Majdi S. Abbas msa at latt.net
Sat Feb 21 02:44:44 UTC 2015


Apologies for the top posting and brevity; I'm on a mobile device.

The next generation of the protocol should include a reference type field that could be used to specify whether or not the refclock field should be decoded as v4, v6, or ASCII (perhaps even UTF-8.)

I realize that doesn't help you today but it's likely that chimes in a v6 world will be larger.

I already have a problem with the loop detection algorithm; I have multiple pool servers slaves to my refclocks.  Some are dual stack, and some are v4 alone because their hosting providers do not support v6 yet.

As a result of this mixed protocol operation, the v4 only and dual stack hosts appear to be slaved to different refclocks when they are not, breaking the loop detection/selection algorithm.

So that's the bad news.  On the bright side a new protocol is an opportunity to fix this and other similar issues that exist now.

--msa

> On Feb 20, 2015, at 17:49, Harlan Stenn <stenn at ntp.org> wrote:
> 
> A number of refclock drivers support the "stratum" directive, which can
> be used to say that the stratum of the current refclock is other than 0,
> the default.
> 
> I've never really understood why this was particularly useful.  But
> that's not the point.
> 
> The other problem is that the REFID has been a 1-4 char ascii string for
> refclocks (stratum 0) and it's treated as an IPv4 address for anything
> other than S0.  So things get confusing if you specify a refclock at
> other than S0 because the refid doesn't decode in an expected way and we
> have no current way to tell downstream clients that they're talking to a
> refclock at other than S0.
> 
> Thoughts or discussions?
> 
> -- 
> Harlan Stenn <stenn at ntp.org>
> http://networktimefoundation.org  - be a member!
> _______________________________________________
> hackers mailing list
> hackers at lists.ntp.org
> http://lists.ntp.org/listinfo/hackers


More information about the hackers mailing list