Re: [TICTOC] minutes for joint meeting of ntp/tictoc at IETF 88

[Alexander Vainshtein <[e-mail]>]

Tal,
I fully agree with you that having two different drafts addressing the same problem is most unfortunate (even if this is neither the first nor, most probably the last time IETF encounters this issue :-).

At the same time I do not see any easy way to merge the two approaches:
- From my point of view the approach taken in the TICTOC WG draft breaks the MPLS data plane architecture which allows only a limited set of "alert mechanisms" for intercepting a labeled packet and manipulating its payload instead of regular forwarding
- The alternative proposal is an attempt to find a solution that, while delivering the same functionality, would stay within the boundaries of this architecture.

Regards,
     Sasha


> -----Original Message-----
> From: Gregory Mirsky [mailto:[e-mail]]
> Sent: Monday, December 09, 2013 7:14 PM
> To: Tal Mizrahi; [e-mail]; [e-mail]; NTP Working Group;
> Shahram Davari ([e-mail]); Alexander Vainshtein; Stewart Bryant
> <[e-mail]> ([e-mail]); John E Drake
> ([e-mail])
> Subject: RE: [TICTOC] minutes for joint meeting of ntp/tictoc at IETF 88
> 
> Adding John.
> 
> -----Original Message-----
> From: Tal Mizrahi [mailto:[e-mail]]
> Sent: Monday, December 09, 2013 7:11 AM
> To: [e-mail]; [e-mail]; NTP Working Group; Shahram Davari
> ([e-mail]); Gregory Mirsky; Alexander Vainshtein; Stewart Bryant
> <[e-mail]> ([e-mail])
> Subject: RE: [TICTOC] minutes for joint meeting of ntp/tictoc at IETF 88
> 
> Hi,
> 
> I apologize for this late comment, as I could not attend the IETF meeting.
> 
> >Greg Mirsky presented his new draft from the MPLS working group on
> >Residence Time Measurement. There were a number of questions related to
> >the technical content of the draft.
> >Discussion should continue on the tictoc and mpls mailing lists.
> >(draft-mirsky-mpls-residence-time-00.txt)
> >
> >The subject of the current draft on Timing over MPLS that is being
> >working in the tictoc working group was addressed.
> >There are still concerns regarding a way forward for this draft. After
> >much discussion, it was decided that the draft would proceed in the
> >tictoc working group as an Experimental RFC.
> 
> It looks like we are heading towards having two drafts that solve the same
> problem, which is most unfortunate.
> Isn't there a way to update Shahram's draft in a way that will satisfy Greg and
> Sasha's concerns and eliminate the need for draft-mirsky-mpls-residence-time?
> For example using GAL instead of a dedicated LSP in Shahram's draft is an
> option that has been discussed before, and may be the least painful
> compromise.
> 
> Honestly, the worst possible outcome is to define two different solutions. We
> obviously don't expect vendors to implement both solutions.
> 
> I believe it is our obligation to reach a single solution for this problem.
> 
> Thanks,
> Tal.
> 
> 
> -----Original Message-----
> From: TICTOC [mailto:[e-mail]] On Behalf Of Karen O'Donoghue
> Sent: Friday, December 06, 2013 11:03 PM
> To: [e-mail]; NTP Working Group
> Subject: [TICTOC] minutes for joint meeting of ntp/tictoc at IETF 88
> 
> Folks,
> 
> Below are the minutes for the joint meeting of tictoc and ntp at IETF 88. If you
> have any questions or comments, feel free to provide them.
> 
> Regards,
> Karen
>