Re: ipv6nd-02.txt & autoconfig

[Albert Manfredi <[e-mail]>]

[e-mail] wrote:
> 
> Disclaimer: This is grumpy response.

I noticed.

> Bert, did you actually digest the paragraph you responded to? SEL field
> _is_ a neat-o feature, but it _is_ used to create logical ATM interfaces.
> End of story.

Did this proclamation get carved in stone? Just how much history do we
have to put with with behind IPv6 over ATM? Is there a big worry about
legacy networks?

What's wrong with saying "Do not use SEL when running IPv6 over ATM," for
instance? "Use of SEL is only for native ATM comms." Or "Use of SEL is
only permitted after the calling party is is made aware of the fact the
the called party is directly reachable over the ATM net." 

Or, alternatively, "All IPv6 communications over ATM SHALL be accomlished
over SEL = x (perhaps x = 0)."

The objective is exceedingly simple: use the 48-bit ESI _only_ to
generate unique IPv6 addresses using autoconfiguration. Do not use SEL.

Are you concerned that the IETF can't make such a ruling? That's fine, I
never implied it should be the IETF. It could be the ATMF. The simple
fact is that autoconfiguration would become much simpler this way, all to
the benefit of public perceptions of IP over ATM and ATM's survival.

Bert
[e-mail]