[IP] the undead urban myth of the LOC/EID split

[David Farber <dave@farber.net>]

Begin forwarded message:

From: mo@ccr.org (Mike O'Dell)
Date: October 29, 2008 8:28:25 PM EDT
To: dave@farber.net
Subject: the undead urban myth of the LOC/EID split


Dave,

an indulgence if you would.

there is a persistent urban myth (which gets repeated here
with some frequency) which states that splitting "network
addresses" into location-dependent and location-independent
components is the secret to life, the universe, and everything.

i know that myth quite well because once upon a time i subscribed
to it and made a serious proposal to do just that with IPv6.

But if you want to find out why the myth is wrong and what it
takes to have it work right from first principles, you're going
to have to read a book that will likely take some work:

	"Patterns in Network Architecture:
	 A Return to Fundamentals"
	 	by John Day

It contains more than a few deeply profound insights.
Among other things, you'll discover why "global addresses" are
an abberation, and that "NAT" is an absolutely natural technique
to use in structure networks - it's just the introduction of an
arbitrary abstraction encapusulation. In fact, the ugliness of
"NAT" is directly related to how, uh, "unfortunate" the
underlying architecture really is.

this is indeed a shameless plug for John's remarkable book.
if you really want to know what a clean, deeply elegant
network architecture based on solid fundamentals
can look like, read his book.

	cheers,
	-mo

Full and Fair Disclosure:
I reviewed the text along the way as a work in progress.







-------------------------------------------