Fred Baker wrote:
The funny thing is that I'm not convinced that this is a change.
As I have said in this and other fora, there is an argument for
functional complexity in the network, and routing is its poster child.
...
What we need to do is figure out how to let the intelligent network core
work cooperatively with the intelligent edge to let it do intelligent
things. Right now, the core and the edge are ships in the night, passing
and occasionally bumping into each other.
Perhaps the requirement is to stop viewing the choices as adversarial and
consider them in terms of potentially complementing each other. The
infrastructure is not a goal; it is there to support end-user capabilities.
(To the folks actually using the Internet, infrastructure is overhead.)
At its very worst, e2e has never said that there should be no intelligence in
the infrastructure, since as you note, routing has been there from the start
and ain't exactly trivial.
What has been observed, however, is that it is better to keep a function out
of the infrastructure, when possible, because it is both easier to adopt and
easier to change, when it is at the edges.
Over time, what is outside the infrastructure often becomes part of it. DNS
is a good example, in terms of application-level utility. In a different
sense, this can even be true within application services. MIME was carefully
designed as an edge service. No change to the mail transfer world was
required. But try running a serious email service without MIME support today.
And that probably suggests the design philosophy: Start be defining a service
as strictly at the edges. Preferably user hosts, but failing that, in edge
networks.
If support is required deeper than that, fine. But at least have a good
understanding of why and try to make it as minimal as possible.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf