Re: [IPv6] Removing extension headers from packets (revisited)

["Pascal Thubert (pthubert)" <[e-mail]>]

Hi Bob: I agree with Ole and Tom; 

as an example, the RPL option is one. It signals the routing topology (think, VRF) and allows for the lazy route maintenance that allows RPL to remain low power.

RFC 9008 allows the option to be injected natively in the packets, meaning that a packet leaving the RPL domain to an application somewhere in the cloud will still have it.

Tom's proposal would be beneficial to the rest of the Internet.

Another on the works: draft-pthubert-detnet-ipv6-hbh allows DetNet signaling in HbH. The alternate on the table right now is to place that info deep in the packet, after the UDP header, and using an MPLS encoding. I'm not sure about which silicon will happily go that deep regardless of whatever header can come first, or what sense it makes to bring MPLS in the party. IPv6 HbH brings a clear value there IMHO.

Cheers,

Pascal
> -----Original Message-----
> From: ipv6 <[e-mail]> On Behalf Of Bob Hinden
> Sent: Tuesday, August 15, 2023 10:38 PM
> To: Ole Troan <otroan=[e-mail]>
> Cc: Bob Hinden <[e-mail]>; Tom Herbert
> <tom=[e-mail]>; 6MAN <[e-mail]>
> Subject: Re: [IPv6] Removing extension headers from packets (revisited)
> 
> Ole,
> 
> > On Aug 14, 2023, at 11:51 PM, Ole Troan
> <otroan=[e-mail]> wrote:
> >
> >
> >> On 14 Aug 2023, at 22:12, Tom Herbert
> <tom=[e-mail]> wrote:
> >>
> >> Could it be acceptable for routers to remove HBH EH or routing header
> >> EH when a packet exits a limited domain? I realize this is contrary
> >> to
> >> RFC8200 and might be considered an architectural abomination, but on
> >> the other hand this might be a practical way to improve the
> >> deployability of extension headers without any material adverse side
> >> effects (other than packets being dropped which might happen anyway
> >> if packets with EH escaped the limited domain).
> >
> > I don’t see anything principally wrong with this approach.
> > If the HBH option is specifically designed as a mechanism for host to
> network in-band signalling.
> 
> Which HBH options are designed for host to network in-band signaling?
> 
> Bob
> 
> 
> >
> > My assumption would be that the signalling was between the host and the
> first-hop router and the first router processing the option would remove
> it. Would that work for your problem?
> > I’m a little more skeptical about removing the header further into the
> network, aka limited domain egress… at least that gets into more
> “definition” issues.
> >
> > O.
> >
> >
> > --------------------------------------------------------------------
> > IETF IPv6 working group mailing list
> > [e-mail]
> > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> > --------------------------------------------------------------------