Re: draft-carpenter-v4v6tran-framework

[Xing Li <xing@cernet.edu.cn>]

<alpine.LRH.2.00.1008301425100.11133@netcore.fi>
<m2vd6r3c5d.wl%randy@psg.com>
In-Reply-To: <m2vd6r3c5d.wl%randy@psg.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
X-AIMC-AUTH: xing
X-AIMC-MAILFROM: xing@cernet.edu.cn
X-AIMC-Msg-ID: 6CAPQuYB
Sender: owner-v6ops@ops.ietf.org
Precedence: bulk
List-ID: <v6ops.ops.ietf.org>


As a person who has been designing and operating CERNET (IPv4) and
CERNET2 (IPv6), I can share my experience. My comments are

(1) As regards CERNET deployment in 1994, we consulted a lot of
references. Our trick was to get the references from the people who had
been doing the real work. Not the second-hand information.

(2) For the CERNET2 development in 2004, we couldn't find much
information, so we had to do our own trail-and-error. We have designed
CERNET2 as a single-stack IPv6-only network. In 2004, some people said
we should run dual stack. But now, a lot of people told us that the
IPv6-only was the right decision. Similar things have happened with
IPv4-over-IPv6 and stateless translation.

So, I applaud this effort and the aspects or dimensions listed in
draft-carpenter-v4v6tran-framework. My point is that the document should
summarize the real work, not only the conceptual discussions.

Regards,

xing



Randy Bush ��:
>> I applaud any and all such work, but I will note that I'm somewhat
>> skeptical of another IETF effort in this front.  The previous
>> experience was a challenge enough.
>>     
>
> those who do not remember ngtrans are doomed to repeat it.
>
>   
>> I wonder if similar goals could be met by working on documenting these
>> in a wiki. Write drafts if it helps focus discussion, but turning the
>> crank to produce 10+ RFCs that will get outdated soon enough doesn't
>> seem like worth the effort.
>>     
>
> agree.  my guess is that documentation is either un-useful pontification
> which lacks detail or is useful but ephemeral.
>
> randy
>
>
>