Hi Jeffry,

In practice, the profile-specified value is merely treated as a default 
(minimum value),  and is configurable on nearly every device I have worked 
with, bar some OEM modules embedded in other products. Higher message rates are 
very useful in enterprise environments with high PDV, where Telecom profiles 
are not the answer. Otherwise agreed, agreed and agreed.

What I know is that there is precedent, namely Brilliant / Juniper GMs which 
supported mixed uc/mc operation, used to ignore 0x7F for unicast delay 
response. While in breach of IEEE 1588, it was the sensible thing to do. But I 
suppose it is a moot point, at least in the light of this profile, just like 
that product line is dead now.

Anyhow, something to be recorded for the future - "useful information" is the 
key.

Best regards,
Wojciech

--

Wojciech Owczarek


  Original Message  
From: [email protected]
Sent: 16 July 2018 6:07 pm
To: [email protected]
Cc: [email protected]; [email protected]; 
[email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [TICTOC] [Ntp] WGLC draft-ietf-tictoc-ptp-enterprise-profile

My understanding of the Enterprise profile is that the logMessageInterval field 
is simply ignored in favor of a profile-specified value. The profile is 
designed to live alongside non-Enterprise nodes on the same PTP network.

All other profiles are going to put 0x7F in the logMessageInterval field on a 
unicast reply, so the field can't be used for auto-discovery of parameters with 
any profile.

I think using 0x7F was a bad idea in the first place, but it's a very clear 
requirement in 1588-2008. I can't think of any good reason to remove useful 
information simply because a message was unicast instead of multicast. There's 
already a flag for that in the header.

Best regards,

Jeffry


---- On Mon, 16 Jul 2018 05:15:31 -0500 Wojciech Owczarek 
<[email protected]> wrote ----

All,

I might have asked this question before and I think I know the answer, but just 
to re-state this:

This being an IEEE 1588 profile, a recommendation to set logMessagePeriod in 
the enterprise profile delay response header to be the configured master 
delayReq period, rather than the 0x7F as per the standard for any unicast 
message, is out of the question?

This would allow the delay message interval to be autodiscovered like it is 
with multicast delayReq/resp. I have seen this done in the field, just like 
some slave implementations will try unicast delayReq first, and use multicast 
if no response received. 0x7F makes sense as an indication that the interval 
should be negotiated, but not where unicast is used without negotiation. This 
only adds an extra variable to the otherwise automatic slave configuration.

I have not been keeping up with the protocol revision work, but does anyone 
know if P1588 was able to tackle this?

Regards,
Wojciech

--

Wojciech Owczarek


  Original Message  
From: [email protected]
Sent: 16 July 2018 9:12 am
To: [email protected]; [email protected]
Cc: [email protected]; [email protected]; [email protected]
Subject: Re: [TICTOC] [Ntp] WGLC draft-ietf-tictoc-ptp-enterprise-profile

Neither do I, thanks a lot, Tal!

 

BR

  Heiko

 

 

-- 

Heiko Gerstung 
Managing Director

MEINBERG® Funkuhren GmbH & Co. KG
Lange Wand 9
D-31812 Bad Pyrmont, Germany
Phone:    +49 (0)5281 9309-404
Fax:        +49 (0)5281 9309-9404

Amtsgericht Hannover 17HRA 100322
Geschäftsführer/Management: Günter Meinberg, Werner Meinberg, Andre Hartmann, 
Heiko Gerstung

 

On 04.07.18, 23:34  "TICTOC im Auftrag von Douglas Arnold" 
<[email protected] im Auftrag von [email protected]> wrote:

 

I have no objections to Tal's suggestions.  

 

--- Doug

 

On Tue, Jul 3, 2018 at 3:27 AM, Tal Mizrahi <[email protected]> wrote:

Hi, 

 

I believe this draft is almost ready for publication, with a couple of comments 
below.

 

A couple of comments:

- The section "Forbidden PTP Options" should include also unicast discovery and 
unicast negotiation (mentioned in previous sections).

- In the "Security Considerations" section - I suggest to add a general 
comment: "General security considerations of time protocols are discussed in 
[RFC7384]".

 

Thanks,

Tal.

 

On Wed, Jun 27, 2018 at 6:35 AM, Karen O'Donoghue <[email protected]> wrote:

Folks,

 

This email begins a quick followup WGLC for the following draft:

https://datatracker.ietf.org/doc/draft-ietf-tictoc-ptp-enterprise-profile/

 

This document has been around for a long time, and it has been aligned with the 
updates planned to IEEE 1588. Please respond on the mailing list with your 
recommendation on publication of this document along with any final comments 
that you may have. This WGLC will close on Wednesday 11 July 2018. 

 

Regards,

Karen

 

 

_______________________________________________
ntp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ntp

 


_______________________________________________
ntp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ntp




 

--

Doug Arnold

Principal Technologist

Meinberg USA

+1-508-309-6268

 

_______________________________________________ TICTOC mailing list 
[email protected] https://www.ietf.org/mailman/listinfo/tictoc
_______________________________________________
TICTOC mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tictoc


_______________________________________________
TICTOC mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tictoc

Reply via email to