The IP Over NBMA (ION) Archive

["Eric W. Gray" <egret@nh.ultranet.com>]

Masataka Ohta wrote:
> 
> Eric;
> 
> > > I can't counter argue agaist a simple statement of "NHRP is important
> > > because short-cutting is useful for unstated admin reasons.".
> 
> > It is precisely this kind of statement that makes me wonder if I'm holding
> > discourse with a human being.  How can you miss the obvious frailty of this
> > statement in light of your previous reponses to technical arguments?
> 
> If you or anyone else doesn't have any technical counter argument,
> I'm fine.
> 
> To summarize the discussion,
> 
>         1) There has been ATMARP.
> 
>         2) NHRP was believed to be applicable to cases where ATMARP
>         is not scalable enough.
>

People believe that NHRP is useful for reasons that are completely orthogonal 
to ATMARP as well. 

> 
>         3) It is proven that NHRP is only as scalable as ATMARP.
>

Not true.  As I said earlier, ATMARP is MARS-scalable with small LISs
which prevent ATMARP from being as scalable as NHRP.

> 
>         4) Thus, NHRP, which is complicated, has no better applicability
>         than ATMARP.
>

In most important ways, NHRP is a relatively simple extension of ATMARP.

> 
>         5) Some people still claimed that NHRP is applicable to cases
>         where a single ATMARP subnet is no good because of some
>         administrative reasons.
>

"Administative" here means non-technical.  Note that arguments were put
forward which were not necessarily "administrative" and that many of the
networks that exist are organized based on "non-technical" reasons.  To
ask real-world people to re-organize their actions based on a particular
"technical" perspective is - well - naive (to be charitable).

It would be very interesting to watch you explain to a hospital staff why
they should organize their network based on your technical considerations
rather than their own!

> 
>         6) It is shown that, with the stated administrative reasons,
>         cut-through is no good and ATMARP is good enough.
>

Shown to whom? Anyone who followed that discussion knows that it did
not conclude that way. You stated that the only reason such a thing
would be necessary would require use of firewalls.  Others disagreed.
That's how it was left (nothing particularly conclusive about it).

> 
> So, the engineering conclusion is that NHRP is unnecessary
> complication of ATMARP and should be dismissed.
>

"Bridge" the gap!  Your conclusions are complete non-sequiturs at each point.

> 
> Feel free to technically counter argue against the conclusion.
> 

Am I the only one in this discussion who feels that this summary
leaves some of the technical arguments out?

>                                                         Masataka Ohta