Re: [Idr] Adoption and IPR call for draft-wang-idr-vpn-prefix-orf-03.txt (8/16 to 8/30)

["[e-mail]" <[e-mail]>]

Hi, All:
I support the adoption of this draft.
The updated contents have addressed main concerns raised during the previous discussions. Some remained comments can be addressed after its adoption via the efforts of WG and the implementation experiences of this mechanism.

This mechanaim is one kind of control plane protection. The influences of the forwarding plane will be limited on the offending routes. Doing so can certainly protect the routes receiving, processing and forwarding of other VPNs from being influenced.



Yue Wang
China Telecom Research Institute
 
From: Robert Raszuk
Date: 2022-08-24 01:32
To: Jeffrey Haas
CC: idr@ietf. org; Zhuangshunwan; Sue Hares
Subject: Re: [Idr] Adoption and IPR call for draft-wang-idr-vpn-prefix-orf-03.txt (8/16 to 8/30)
Hi Jeff,
 
> First when BGP max prefix used session should get terminated with a peer. 

That's not their intent.

Of course. This was just along the lines of their comparison to max prefix limit. 
 
They're doing this on VPN routes, not iBGP distributed IPv4 unicast routes.

Of course. 

But as you know for VPNs there are few different types of demux VPN labels. One of them is per vrf label where before any forwarding IPv4 unicast lookup is done at the VRF's FIB. 

If you have scenarios where you think dropping the VPN routes that would be used in a L3VPN VRF context is harmful, I'm sure the authors would appreciate explicit examples.

Few examples:

*  EIBGP load balancing will not work for multihomed site and per VRF label. 
* PE-CE protection will not work. 
* Active-Backup multihomed site breaks if we start dropping say more specific routes and suddenly large volume of traffic will be taking different then planned forwarding path 
etc ... 

Many thx,
R.