Re: An idea: GxSE

[Jun-ichiro itojun Hagino <itojun@iijlab.net>]

>I understand how you feel, however, I feel rather strongly that we should
>do this the right way from the start.  One of the biggest problems with
>NAT is that it was an afterthought, thus broke many things.  We're not
>designing a protocol here, just ways to handle certain situations.

	i disagree about the "from the start" part.  the deployment is already
	started here and there.  maybe this is because i'm in Japan, but
	i feel it way too late to make any changes to the way routers/hosts
	work.

	also, remember why GSE proposal was not integrated into IPv6 - 
	IPv6 is designed based on IPv4 experiences, and tries to benefit
	from the success of IPv4.  if we add too many fundamental design
	changes into IPv6 (like 8+8 address rewrites on routers)
	we may not be able to success in IPv6, like IPv4 did.

>Furthermore, we're discussing technology in its infancy.  There is no
>reason not to explore the possibilities.  Remember, WE make the rules, so
>we might as well look at all the options to the good ones.  These
>decisions will be with us for at LEAST a decade, let's make decisions we
>can deal with.

	well, IPng effort was started year 1992 and now it's year 2001.
	i don't really agree with the "infancy" argument.
	you may notice that the email have went through IPv6-to-IPv4
	translator to the mail list server (it may show the evidence on
	Received: lines).

itojun