Re: [multrans] Issues charts for presentation at the BoF

[Tom Taylor <[e-mail]>]

OK, thanks.

On 28/06/2011 1:31 PM, Dan Wing wrote:
>> -----Original Message-----
>> From: Tina Tsou [mailto:[e-mail]]
>> Sent: Tuesday, June 28, 2011 10:13 AM
>> To: 'Dan Wing'; 'Tom Taylor'; 'Jacni Qin'
>> Cc: 'Multicast Transition'
>> Subject: RE: [multrans] Issues charts for presentation at the BoF
>>
>>
>>
>>
>> We keep our promises with one another – no matter what!
>>
>> Best Regards,
>> Tina TSOU
>> http://tinatsou.weebly.com/contact.html
>>
>> -----Original Message-----
>> From: [e-mail] [mailto:[e-mail]] On
>> Behalf Of Dan Wing
>> Sent: Tuesday, June 28, 2011 9:58 AM
>> To: 'Tom Taylor'; 'Jacni Qin'
>> Cc: 'Multicast Transition'
>> Subject: Re: [multrans] Issues charts for presentation at the BoF
>>
>>> -----Original Message-----
>>> From: [e-mail] [mailto:[e-mail]] On
>>> Behalf Of Tom Taylor
>>> Sent: Tuesday, June 28, 2011 5:54 AM
>>> To: Jacni Qin
>>> Cc: Multicast Transition
>>> Subject: Re: [multrans] Issues charts for presentation at the BoF
>>>
>>> Thank you for your comments. Responses below.
>>>
>>> On 27/06/2011 10:45 PM, Jacni Qin wrote:
>>>> hi Tom,
>>>>
>>>> Some comments,
>>>>
>>>> Page 8:
>>>> Issue: translation of a given address tends to happen
>>>> at multiple nodes, and at different stages within the
>>>> same node.
>>>> ● Need the same mapping or its inverse, as applicable, to be
>>>> used each time a given address is translated. Implies inter-
>>>> nodal coordination in some form.
>>>>
>>>> Jacni>: The coordination is probably required in the stateful
>>>> translation. While in some stateless cases, maybe not.
>>>
>>> [PTT] I agree there is less coordination for stateless, but at least
>>> the
>>> different nodes have to be provisioned with the special prefix being
>>> used to construct IPv4-embedded IPv6 addresses in this network.
>>
>> They need not necessarily be 'provisioned' (configured).  They could
>> learn the network's prefix for multicast translation.
>> draft-ietf-behave-nat64-learn-analysis discussed a bunch of techniques
>> for learning unicast NAT64; I expect almost all of those could be
>> useful for multicast NAT64 (or multicast NAT46) and have pros/cons.
>> [Tina: So what changes do you want to see from Page 8?]
>
>
> how about:
>
> OLD:
> Issue: translation of a given address tends to happen
> at multiple nodes, and at different stages within the
> same node.
> ● Need the same mapping or its inverse, as applicable, to be
> used each time a given address is translated. Implies inter-
> nodal coordination in some form.
>
> NEW:
> Issue: translation occurs in multiple nodes, and different
> stages within a single node
> ● Stateful translation:  nodes need to coordinate
> ● Stateless translation:  nodes need same transformation algorithm
>
>
> (overall, in my opinion, the entire presentation could benefit from a 50% reduction of its word count.  This makes a better presentation because the audience won't be tempted to read the slides, and the speaker won't be tempted to read the slides to the audience.  Instead, the speaker provides additional detail to the main points on the slide.).
>
> -d
>
...