I assume you are looking for some sort of quality requirements like the
telephone industry has for their various stratum levels.  I don't think you
will find it.

Basically, stratum 0 is anything from outside NTP.  The "0" doesn't promise
any quality.

There is a mechanism within ntpd for the driver to report a quality estimate.

--
The suespammers.org mail server is located in California.  So are all my
other mailboxes.  Please do not send unsolicited bulk e-mail or unsolicited
commercial e-mail to my suespammers.org address or any of my other addresses.
These are my opinions, not necessarily my employer's.  I hate spam.

The Stratum numbers in the Telecom Sync industry and those used in NTP
hierarchy are different beasts. In Telecom Sync, Stratum 1, 2, 3 et al are
internationally defined levels of frequency accuracy. In NTP Stratum 1
merely means the top of the hierarchical pyramid with Stratum 2, 3 etc below
it. Stratum 0 has been used in discussions I have seen from people at Datum
and TrueTime to represent a GPS/NTP reference, but realistically I believe
that it should be labelled as Stratum 1.

Would  humbly suggest you take a look at www.ntp.org for more information.

Rob Kimberley

-----Original Message-----
From: time-nuts-bounces@febo.com [mailto:time-nuts-bounces@febo.com] On
Behalf Of Hal Murray
Sent: 14 January 2006 18:35
To: albertazzi.mich@tiscali.it; Discussion of precise time and frequency
measurement
Subject: Re: [time-nuts] NTP stratum 0

I assume you are looking for some sort of quality requirements like the
telephone industry has for their various stratum levels.  I don't think you
will find it.

Basically, stratum 0 is anything from outside NTP.  The "0" doesn't promise
any quality.

There is a mechanism within ntpd for the driver to report a quality
estimate.

--
The suespammers.org mail server is located in California.  So are all my
other mailboxes.  Please do not send unsolicited bulk e-mail or unsolicited
commercial e-mail to my suespammers.org address or any of my other
addresses.
These are my opinions, not necessarily my employer's.  I hate spam.

time-nuts mailing list
time-nuts@febo.com
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts

"Rob Kimberley" time.bandit@btinternet.com wrote:

I think that "Stratum 0" has at least two meanings in the NTP context:

A. Part of a KOD message in the protocol.

B. What "ntpq -c peers" says about a locally connected refclock.

(B) isn't unique to Datum/TrueTime GPS's, it's for all refclocks with ntpd
drivers.

Tim.

From: "Rob Kimberley" time.bandit@btinternet.com
Subject: Re: [time-nuts] NTP stratum 0
Date: Sat, 14 Jan 2006 18:48:16 -0000
Message-ID: !&!AAAAAAAAAAAYAAAAAAAAAOYAZyOzV8ERq+LmT45ypI7CgAAAEAAAAAzy/YBW0+RBmivoTTPkkh0BAAAAAA==@btinternet.com

I agree, they are completely different beasts.

Um, no. They are nationally specified in the USA. You find them in ANSI T1.101.
For their international variants you have to go to ETSI EN 300 462 series, as
well as the ITU-T Rec. G.810 to G.813. Neither the ETSI or ITU-T standards uses
Stratum indications, instead they use designations such as PRC, SSU and SEC.

Indeed. It is really just a reference chain length from whatever the reference
is, which may anything form a Ceasium-beam with coordinated time to an
undiciplined XO.

Indeed.

Also, the references from earlier this thread adds up. The RFC 1305 is as such
a very interesting article.

Cheers,
Magnus

Magnus Danielson wrote:

Yes the T1.101 is a USA standard, but when it was finalized there was
international contributions and every
attempt was made to harmonize the international standards and the USA
T1.101. I was active in the T1X3
meetings.

Bill K7NOM

From: Bill Janssen billj@ieee.org
Subject: Re: [time-nuts] NTP stratum 0
Date: Sat, 14 Jan 2006 16:35:55 -0800
Message-ID: 43C998EB.2030605@ieee.org

Bill,

I think you are missing the point. No question about from where the material
came, but I was commenting on the outcome and if you read the resulting
documents you will find that these terms did not make it into the international
standard, which was my point.

ETSI changes the specs for Stratum-3/SEC clocks and Stratum-2/SSU clocks to
match the needs for the European PDH and SDH needs. That is the "2048 kHz
based" networks also known as "Option I SDH" where as the ANSI specs for the
clocks is in the "1544 kHz based" networks also known as "Option II SDH" (There
is also an "Option III SDH" but lets just leave that in obscurity shall we?
It's 64 kHz based.). The ITU-T Rec. G.781 only uses the Stratum 2, 3 and 3E
terms to reference to SONET clocks but their specs is refered to as G.812
Type II, III and IV. After the introduction, they are only refered to as
QL-ST2, QL-ST3E and QL-ST3 (in falling quality order). Thus, they are refered
to but not the primary term.

The Stratum 3E (Stratum 3 Enhanced) and Stratum 4E (Stratum 4 Enhanced)
indicates that it is not good to try to assign a strict number series of
integers since you might need to insert something into the series and they did.
Ah well.

So, the use of Stratum for clock quality is something out of the Bellcore (now
Telcordia) and ANSI specs. Also, in these they not only use Stratum. The SMC
(SONET Minimum Clock) is added in ANSI T1.105.09 and the TNC (Transit Node
Clock) is added in ANSI T1.101 (Annex A).

So, trying to get a common picture by combing the Stratum numbers from T1.101
and from RFC 1305 will just result in computation error.

We are thankfull for the T1X3 committe and the work done there, but what popped
out on the other end of ETSI and ITU-T is a bit different, that is all I'm
saying.

Cheers,
Magnus