As far as intra-AS infrastructure links are concerned, I'm perfectly fine with 
the draft.

HOWEVER, EBGP is a huge can of worms. While we can definitely make EBGP work 
over link-local addresses, the interface name becomes part of EBGP neighbor ID 
(at least in Cisco IOS), leading to a nightmare scenario if you have to move 
the peering (or worse: IXP) link to another interface in hurry. At least Junos 
has configuration identifier renaming functionality ...

Hallway discussions with a few IXP operators and IXP members during last week's 
RIPE65 also indicated some lack of enthusiasm for this idea. I know the sample 
was statistically irrelevant, but it was clear not everyone wholeheartedly 
embraces EBGP-over-LLA concept.

Ivan

> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On Behalf Of
> Benoit Claise
> Sent: Monday, October 01, 2012 11:11 AM
> To: [email protected]
> Subject: [OPSEC] Feedback on draft-ietf-opsec-lla-only-01
> 
> Dear all,
> 
> I discussed one topic off line with Eric, regarding
> http://tools.ietf.org/html/draft-ietf-opsec-lla-only-01
> For the sake of openness, here it is again on the list.
> 
> One on hand, the draft mentions:
> 
>     Lower configuration complexity: LLAs require no specific
>     configuration, thereby lowering the complexity and size of router
>     configurations.  This also reduces the likelihood of configuration
>     mistakes.
> 
> On the other hand, the draft mentions:
> 
>     These link-local addresses SHOULD be hard-coded to prevent the change
>     of EUI-64 addresses when changing of MAC address (such as after
>     changing a network interface card).
> 
> So the question is: who is going to configure this? If the NMS, there is
> not much of a gain in term of lower configuration complexity.
> 
> We discussed the possibility of a global config on the router, for which
> the link-local addresses for that router would be generated from the UUID.
> 
> Regards, Benoit
> 
> 
> _______________________________________________
> OPSEC mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/opsec

_______________________________________________
OPSEC mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsec

Reply via email to