Re: [Status] Jari Arkko's BLOCK on charter-ietf-spring-00-06

[Robert Raszuk <[e-mail]>]

Hi Hannes,

> | Yes it will be allowed to get into any network, however it is very
> | easy not to examine extension headers on such packets hence only
> | provide v6 transit as any DFZ would provide today without any SR
> | enabled.
>
> so you are advovating that internet DFZ router MUST NOT support IPv6-SR
> based forwarding as a consequence of your statement above ?

No .. not at all. DFZ routers will first check if the v6 header
belongs to infrastructure subnets. (Those will be filtered on the
edges perhaps with the exception of some heavy rate limited ICMPs - in
fact I am not sure if ICMP should support SR at all).

If the v6 header belongs to infrastructure range SID lookup will
follow if not outer v6 header will be used for forwarding (as today).

/* Now I am expecting few emails with complains ..Ohhh my ASICs can
not choose switching vector based on the outer IP header address at
line rate ;) */

> i am not sure those things can get fixed at a architectual
> level at all, given the amount of failed attempts so far.

Do we have a list or wiki page of those "failed attempts" ? In fact so
far I think the focus has been on the charter discussion which is
quite orthogonal to the SR architecture ..

Best,
r.