Re: [PWE3] Working group last call and IPR call for dynamic Multi-Segment PW drafts

[Wen Lin <[e-mail]>]

Hi Matthew,

Again, thank you for the reply.

As I have stated before that the reason we shall have a route target associated with the MS-PW NLRI is that it is a VPN route - i.e. the MS-PW NRLI is advertised by the T-PEs for the Layer 2 VPN.  

The latest version of this dynamic Multi-Segment PW draft took out the section addressed the "S-PE addressing" requirement and also the reference to the explicit route TLV(explicit path). This makes the proposal a more generic approach for achieving dynamic setup of MS-PW using type 2 AII. I still think a route target should be associated with the MS-PW NLRI. Or is there any bad effect for associating route target in this case?

Thanks,
Wen 



-----Original Message-----
From: Bocci, Matthew (Matthew) [mailto:[e-mail]] 
Sent: Monday, October 08, 2012 10:43 AM
To: Wen Lin; [e-mail]
Subject: Re: [PWE3] Working group last call and IPR call for dynamic Multi-Segment PW drafts

Hi Wen,

I think the scenario that you are describing is when you are using AII type 2 for auto-discovery. This draft does not exclude doing this, but that's a different scenario. Instead, it describes a scenario where we are using an aggregated AII type 2 to distribute routes among a known set of S-Pes and T-Pes.

If you also want to use AII type 2 in auto-discovery scenarios, when I think this needs to be documented in a separate draft. However, I can add a sentence to the dynamic MS-PW draft to state that a RT MAY be included along with the NLRI.

Matthew 

On 04/10/2012 04:07, "Wen Lin" <[e-mail]> wrote:

>Hi Matthew,
>
>Thank you for the reply.
>
>On the T-PE, the MP-BGP advertises this MS-PW NLRI for the VPWS 
>instance and the layer 2 VPN. A MS-PW may go across different ASes.  
>The route target serves as a VPN identifier and helps to control the route
>distribution.   
>
>Also we shall not exclude using the same MS-PW NRLI for the 
>auto-discovery of S-PE or T-PE by using the type 2 AII.
>We shall also not exclude using the same NLRI for the VPLS. The route 
>target will help us identify the VPLS domain that the VSI belongs to 
>and helps us build different topologies for the PW.
>
>
>Thanks,
>Wen
>
>
>-----Original Message-----
>From: Bocci, Matthew (Matthew) 
>[mailto:[e-mail]]
>Sent: Tuesday, October 02, 2012 5:25 AM
>To: Wen Lin; [e-mail]
>Subject: Re: [PWE3] Working group last call and IPR call for dynamic 
>Multi-Segment PW drafts
>
>Wen,
>
>Thanks for your comment. I am not sure why sending or processing a 
>route target would be mandatory in the case of MS-PW routing. This is 
>somewhat different from the auto-discovery case described in RFC6074. 
>In MS-PW routing, you know a-priori the PEs that can participate in 
>MS-PW routing and can terminate MS-PWs, the set of PSN tunnel between 
>T-Pes/S-Pes, and the T-LDP next-hop.
>
>Regards
>
>Matthew   
>
>On 05/09/2012 20:40, "Wen Lin" <[e-mail]> wrote:
>
>>When BGP is used for advertising PW address information, I think we 
>>shall specify in the section 6.1.3 that it MUST have one or more route 
>>targets associated with it since the NRLI is for L2VPN with a an AFI 
>>equals to L2VPN(25).
>>
>>Thanks,
>>Wen
>>
>>
>>
>>_______________________________________________
>>pwe3 mailing list
>>[e-mail]
>>https://www.ietf.org/mailman/listinfo/pwe3
>
>
>