Re: [mpls] Fast Reroute for Node Protection in LDP-based LSPs

Isn't the point here to decouple the end point calculation and 
collection of axillary data from the method of  path instantiation?

Given that you know the repair endpoints and the additional information 
needed to forward the packet from the repair endpoint how you 
instantiate the repair is another matter.

Whether an operator wants to use RSVP or SR is up to them.

Something that I would note is that RSVP and strict source routing allow 
the operator to pick any path of their choosing, but as far as I can see 
the loose source routing option requires the complexity of the 
calculation of a path that is congruent both before and after 
convergence to prevent the repair path falling apart during convergence.

- Stewart

On 09/12/2016 19:15, Santosh Esale wrote:
> Hi Stefano,
>             The main purpose of the solution is to provide
> topology independent local protection using RSVP-TE in widely
> deployed LDP based MPLS networks. Link protection is already
> deployed using manually configured RSVP-TE one-hop LSPs. This
> draft addresses node protection.
>
>   
> Now, the solution can also be used to protect segment routing
> hop-by-hop node segments and we would add a note about it.
> Of course, the other solution that you mentioned may fit
> segment-routing too.
>
> Thanks,
> Santosh
> On 12/7/16, 1:30 AM, "Stefano Previdi (sprevidi)" <sprevidi@cisco.com>
> wrote:
>
>>> On Dec 6, 2016, at 11:55 PM, Santosh Esale <sesale@juniper.net> wrote:
>>>
>>> Hi Hannes,
>>>           Good point! We will generalize the solution to cover
>>> segment-routing (SR) too in the next - 05 revision.
>>
>> I don’t think your solution brings anything better than what already
>> covered in ti-lfa draft for segment-routing. See
>> draft-francois-rtgwg-segment-routing-ti-lfa-02.
>>
>> s.
>>
>>
>>> Thanks,
>>> Santosh
>>>
>>> On 12/6/16, 3:17 AM, "Hannes Gredler" <hannes@gredler.at> wrote:
>>>
>>>> hi santosh,
>>>>
>>>> just curious why the proposed solution is constrained to only use LDP ?
>>>>
>>>> IMO what you have suggested here would just fit nicely for protecting
>>>> segment-routing node labels as well. segment routing node-labels are
>>>> "calculated" in a similar fashion than LDP labels as such my guess
>>>> would
>>>> be that this solution applies to SR node labels as well.
>>>>
>>>> suggest to generalize it to:
>>>> "Fast Reroute for Node Protection in hop-by-hop based LSPs"
>>>>
>>>> thanks,
>>>>
>>>> /hannes
>>>>
>>>>
>>>> On 12/5/16 20:59, Santosh Esale wrote:
>>>>> Hello Everyone,
>>>>>                            We have presented the draft
>>>>> - draft-esale-mpls-ldp-node-frr – in
>>>>> MPLS working group in three IETF meetings including the latest one at
>>>>> Seul.
>>>>> However, as the draft is of interest to routing working too, we are
>>>>> initiating this
>>>>> thread to solicit feedbacks from the routing working group. Please let
>>>>> us know
>>>>> your comments.
>>>>>
>>>>> Presentations -
>>>>> https://www.ietf.org/proceedings/93/slides/slides-93-mpls-3.pdf
>>>>> https://www.ietf.org/proceedings/94/slides/slides-94-mpls-3.pptx
>>>>>
>>>>>
>>>>> https://www.ietf.org/proceedings/97/slides/slides-97-mpls-08-ti-frr-iet
>>>>> f-
>>>>> 97-00.pptx
>>>>>
>>>>> Thanks,
>>>>> Santosh (on behalf of authors)
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> rtgwg mailing list
>>>>> rtgwg@ietf.org
>>>>> https://www.ietf.org/mailman/listinfo/rtgwg
>>>>>
>>> _______________________________________________
>>> rtgwg mailing list
>>> rtgwg@ietf.org
>>> https://www.ietf.org/mailman/listinfo/rtgwg
>
> _______________________________________________
> mpls mailing list
> mpls@ietf.org
> https://www.ietf.org/mailman/listinfo/mpls