Hi Scott,

You wrote:

>> None of the A2s fit APT exactly.
> 
> Let's get away from trying to compare entire _packages_ and instead
> look at functional components.  Whether APT fits a particular approach
> isn't as important as whether APT's pieces do.  When we know which
> approaches we prefer, we select building blocks to put them together.
> Let's not get trapped arguing about packages.

While I think it is interesting to break down a proposal into it's
various distinctive functional blocks - especially when contemplating
how to improve this or other proposals - ultimately, I believe it
will be complete, carefully designed, proposals which have the best
chance of solving the scaling problem.

It is no good trying to build a car from a collection of GM and Ford
parts - they will never fit together as well as a car made with a
unified design, so all the components are combined to maximise
synergy between the various sub-systems.

I think it is valuable that Bill's page attempts to identify
commonalities and distinctions between the various approaches to
designing a new architecture.

An important test of Bill's taxonomy of architectural design
approaches is whether all the major, real, live proposals actually
fit clearly within it.

If Bill's page isn't updated to the point where the designers of APT
and ILNP are happy that there is a place in the taxonomy for their
designs, then I think the taxonomy would be only a loose and somewhat
misleading account of the actual design choices which are available.

I think it would be best if the designers suggest changes to Bill's
taxonomy - specific words to modify the current Strategies, or to
create new ones.  I did this with Ivip, and Bill used a subset of
what I suggested.

While the taxonomy doesn't consider a number of architectural aspects
which I think are important distinctions between Ivip and other
proposals, at least Ivip has a place in Bill's taxonomy, as does
LISP-ALT, TRRP and LISP-NERD.

 - Robin

_______________________________________________
rrg mailing list
[email protected]
https://www.irtf.org/mailman/listinfo/rrg

Reply via email to