Re: [spring] I-D Action: draft-previdi-spring-problem-statement-02.txt

["Stefano Previdi (sprevidi)" <[e-mail]>]

Hi Yakov,

On Apr 4, 2014, at 10:53 PM, Yakov Rekhter wrote:
> Stefano,
> 
>> All,
>> 
>> this is the new version of the problem-statement draft incorporating 
>> the latest comments.
> 
> Section 5.1.1.2 is clearly an improvement over what was in the
> previous version, although some text is still problematic.
> Specifically, "C only installs the path via AS2 in its RIB." If it
> is the Adj-RIB-Out, then C wouldn't be able to advertise the path
> via AS3 inside AS1 (as this path is not in the Adj-RIB-Out).  And
> if it is the Loc-RIB, then the Adj-RIB-Out would have only the path
> via AS2 (as the Adj-RIB-Out is populated from the Loc-RIB), and C
> would not be able to advertise the path via AS3 inside AS1. Either
> way, this would contradict the assumption that "C propagates all
> the paths to Z within AS1 (add-path)."


I'll address your BGP concerns on a separate thread.


> You addressed one part of my comments on section 3.


it's already good to hear/read that ;-)


> However,
> the other part of my comments on section 3 is still not addressed.
> Specifically, section 3 of draft-previdi-spring-problem-statement-02 
> still contains the following:
> 
>   The source-based routing model, applied to the MPLS dataplane, offers
>   the ability to tunnel services (VPN, VPLS, VPWS) from an ingress PE
>   to an egress PE, with or without the expression of an explicit path
>   and without requiring forwarding plane or control plane state in
>   intermediate nodes.
> 
> The above claim about the ability to tunnel services is misleading,
> as VPN and VPLS services may use p2mp or (in the case of VPN) mp2mp
> LSPs, and providing any capability for setting up such LSPs is not
> in the SPRING charter.
> 
> With this in mind I propose to replace the above with the following:
> 
>   The source-based routing model, applied to the MPLS dataplane, offers
>   the ability to tunnel services (VPN, VPLS, VPWS) from an ingress PE
>   to an egress PE, with or without the expression of an explicit path
>   and without requiring forwarding plane or control plane state in
>   intermediate nodes, but only if such tunnels are neither p2mp
>   nor mp2mp tunnels.


fine by me, I'll incorporate your comments. Thanks.

s.



> 
> Yakov.
>