[sig-ipv6] BOUNCE sig-ipv6@lists.apnic.net: Non-member submission from [Kenji Rikitake <kenji@k2r.org>]

[owner-sig-ipv6@lists.apnic.net]

>From owner-sig-ipv6  Thu Jul  6 19:08:47 2000
Received: from limmax.k2r.org (limmax.k2r.org [210.160.188.69])
	by whois.apnic.net (8.9.3/8.9.3) with SMTP id TAA121280
	for <sig-ipv6@apnic.net>; Thu, 6 Jul 2000 19:08:45 +1000 (EST)
Received: (qmail 88384 invoked by uid 1000); 6 Jul 2000 09:09:00 -0000
Message-ID: <20000706090900.88383.qmail@k2r.org>
Date: Thu, 6 Jul 2000 18:08:38 +0900 (JST)
From: Kenji Rikitake <kenji@k2r.org>
To: sig-ipv6@apnic.net
cc: Kenji Rikitake <kenji@k2r.org>
Subject: On IPv6 Policy Document Revision
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

I've just browsed the request-for-comment message from APNIC on the
revision of one-size-fits-all principle in RFC2374.  The following
represents my comment on this issue.

I think introducing variable-length prefix on IPv6 will generate another
administrative and political hassles currently going on the IPv4 address
assignment, let alone the multi-homed routing issue.  I also think 
reducing the administrative overhead and supporting easy migration between
different TLAs is one of the key goal of the IPv6 address assignment.

I also think dividing the sites into two /48 and /56 categories will
fall into another administrative and political hassles too, since /56
mask will only allow 255 subnets *AS THE MAXIMUM VALUE*, and extending
/56 to a bigger space will require a total reassignment of /48 space or
using a variable mask, which will result in an even complicated routing.

On the other hand, regarding the population growth rate of the human
being, we will require addresses for no more than 2^37 (1.37 trillion)
people.  This still means 2^11 or 2048 addresses per capita.  While I do
not disregard the possibility of the address assignment to the
extra-terrestrial beings, 2^48 identifiers are so far sufficient for the
human being.

So among the three alternative principles APNIC has proposed, I think

> 1) /64 for single devices (such as mobile phones), /48 for all other sites

is the most preferable choice.

Acknowledgement to Akira Kato, Takashi Arano, Yoshihiro Obata, and Ho
Seongmyong of JANOG on various inputs for this comment.

// Kenji Rikitake

*              sig-ipv6:  APNIC SIG on IPv6 technology and policy issues           *
* To unsubscribe: send "unsubscribe" to sig-ipv6-request@apnic.net *