Re: IPv6 tranisition issues

----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Thursday, December 26, 2002 4:04 AM
Subject: (ngtrans) IPv6 tranisition issues


> This may be slightly OTT for this mailing-list. I am looking for good
> paper/presentations from oponents of IPv6 transition (and perhaps the
> IPv6 answers to these). We are planning to write a paper to form an IPv=
6
> steering committee here in Malaysia and need to be prepared.
>
> Any help would be greatly appreciated.
>
> Some of the points I have noted and gathered:
>
> +Network built on NAT is fine and there is no real need to look at IPv6.
> +What is so special about IPv6 that I cannot patch and do with IPv4.
> +Cannot see the immediate =91cost versus benefit=92 of implementing IPv=
6.
> +Limited number of applications/hardware that are IPv6 enabled.
> +Lack of educational material from service providers, equipment vendors=
,
> and R&E institutions.
> +Poor government/regulatory support and awareness.
> +IPv6 is about 90% complete work still needed on PPP, DNS standards,
> multicast/anycast, etc.
> +Network renumbering, service disruption and other operational issues.
> +Initial =91land-rush=92 for IPv4 address space has resulted in a sligh=
tly
> complacent outlook by US hardware/software developers.
> +No 'Y2K flag-day' for IPv4 address exhaustion. Come on, are we really
> running out of IPv4 addresses? 2^32 is quite a lot of addresses.
>
> Best regards,
> -nick/
>
>
>
There are a lot of very good reasons for going with IPv6:

1. IPv4 and earlier protocols were theoretically experimental and never
really intended for comercial use. The number of IPv4 addresses was never
adequate because there were 4 billion people in the world 1980.

2. Network Address Translators only defer the problem. What a NAT does is
cause a LAN or other subnet to pretend that it is a single timesharing
machine with only 1 global IPv4 address. See a draft protocol named Tered=
o
as one proposal for patching NAT boxes into IPv6. Also take a look at RFC
3056 and the Isatap draft. What NAT has done is to turn the port number i=
nto
a 16-bit extension of an IPv4 address which creates a 48-bit or 80-bit ca=
n
of worms depending on how you look at it.

3. Some of the early participants in what was then known as Arpanet were
each allocated 16 million+ global addresses at essentially zero cost. So
far, only AT&T is one of the few companies that has realized that they ca=
n
make $#!tloads of money by renting out their excess IPv4 addresses at say=
 $7
per month each. As a result, we have one group of users that pay essentia=
lly
nothing for their addresses and another group that pays through the nose.

4. In the usual scheme, the IPv6 address for an individual or cell phone
will consist of a left half (the /64 prefix) that is determined by the
pathway to the device and a \64 suffix in the right half that is
theoretically unique to the device. As a result, changing ISPs or roaming
from one cellular tower to another only involves changing the left half o=
f
the address.
In some instances a large company would get a /48 or /56 prefix and the
companies administrators will complete the /64 prefix. Routing tables wou=
ld
also be a lot easier because the left half is divided into heirarchical
fields.

Michael Cole, [email protected]