Re: DHCP vs. RA... again.

[Jari Arkko <jari.arkko@piuha.net>]

Tony,

> Other deployed networks have hands-on experience showing RA is cheaper for
> them to operate, and the Chicago network was a showcase for the operational
> realities that people experience every day. I am arguing there are cases for
> both deployment models, so we should not be wasting energy debating between
> them, we just need to fix them both so we don't hang in this bizarre state
> where both are required because neither are sufficient as-is. 
>
> If the IESG is going to insist that there has to be a one-size-fits-all
> answer, they all need to be shot during this nomcom round.
>   

If you meant me, I'm already wearing my bullet proof vest :-)

But seriously, I don't like words being put into my mouth.
I said we (the IETF) need to ask hard questions about things
like interoperability, and the case for actual need. But I
also said that we sometimes do have multiple solutions -
e.g., when there's clear need, community consensus that they
are needed, etc. One look at the IETF's WGs should convince
you that we tolerate quite a bit of multiplicity. And I also
believe the IESG's job is not to prevent the well-informed
consensus of the community from going forward unless
there's a clear technical problem.

Having said that, the desire for one solution is indeed
one of our architectural principles (RFC 1958). But
both that RFC, at least this AD, and I believe IETF
folks in general prefer applying common sense over
rigid rules. That is, not something to use to reject
improvements or prevent real needs from being
filled. Of course, this means that a good case has
to be made.

Jari