Re: Minneapolis - agenda items please.

[Johan Ihren <johani@autonomica.se>]

Randy Bush <randy@psg.com> writes:

> [ excuse the metaphoric 'you ]

I didn't see that comment on the first read-through... but fortunately
I did see it before replying ;-)

> > wondering whether it is really so much better to try to cater to
> > all the semi-reachable stuff (nodes that are reachable from
> > certain vantage points but not all) through the complexities of
> > split-DNS everywhere.

> of course the simple answer is one probably should not build such
> sickness in the first place.  but if you insist on such a disease,
> why should i pay for your being lazy and doing a half-assed job?
> 
> > And when we take into accout the increasing numbers of mobile
> > hosts that *change* their vantage point over time it can be argued
> > that it is better to get off the plane and find the same Internet,
> > but with somewhat changed reachability characteristics, than is is
> > to find a different Internet because you're on a different side of
> > a split-DNS gateway point.

> won't the mobile host either be
>   o tunneled to 'inside' and hence will have the inside view of the
>     dns
>   o or living outside and hence have an outside view of the dns?

Yes. And those two are *different*, which is exactly my point. They
shouldn't be. If you and I sit down in two chairs next to each other I
see a certain benefit to us being able to share a common view of the
namespace we call the Internet.

Sure, lots of web sites will be down, or unreachable, whatever. But
that's a property that is inherent to any distributed system: parts
may fail or loose contact. Temporarily or permanently. But the
namespace that describes the parts should always be the same.

> > Split-DNS is not a general solution to be advocated
> 
> similarly to the designs it is trying to accommodate.  but, if you
> insist on those kinds of designs, you need to do the full job.  don't
> complain if insisting on kludge-A leads you to needing kludge-B.  "Oh,
> what a tangled web we weave/when first we practice to deceive."
> 
> > it is a general problem to be avoided. Yes, there are specific cases
> > when it is the least evil choice, but I really do not want to see it
> > as a general solution to this type of problem.
> 
> i doubt any of us like it.  but if you're gonna make a mess, it seems
> your responsibility to contain it.  entropy and all that.

While I agree to the principle (you and I can share the same horse in
the argument against the firewalling brigade) the question I have is
what constitutes the greater evil to the Internet: 

        * a namespace that changes (to the point of being different
          for two people sitting next to each other), but with all the
          stuff you do see reachable 

        * a namespace that is static (in the sense that you and I can
          share it, regardless of whether we're sitting next to each
          other or on opposite sides of the Atlantic), but with some
          stuff unreachable to one or both or us.

It can well be that the first one is better after all, but it is not
obvious to to me that this is the case.

I do not like the semi-reachable stuff myself, but I do realize that
such things are already very common, and are still increasing.

Regards,

Johan