Re: [v4v6interim] [46translation] [BEHAVE] "My" [Re: Proposal for new BEHAVE charter]

[Mark Townsley <[e-mail]>]

The point of "My" was to refer to someone's network, likely other than 
the whole of the IPv4 or IPv6 Internet. Someone will have to purchase, 
setup, and manage the translator understanding well what the demands on 
the translator will be based on the networks being connected. That is 
still understandably vague, but is about as far as we could get during 
the meeting in Montreal in terms of specifics.

At the start of the meeting, Jari and I tried to identify scenarios 
based on direct feedback from individuals which focused the problem a 
bit more... So, not just *any* type of network, but a rather specific 
one by comparison. Those that came to us with  scenarios were quite 
happy to see them portrayed, others that did not expressed either (1) 
confusion of what the scenario was, (2) doubt that the scenario was 
important, or (3) concern that it meant that their scenario, whether 
topologically similar or not, would not be considered by the IETF. So 
this tactic, frankly, didn't work all that well other than to teach us 
that constraining this problem space into specific types of networks is 
very difficult without losing a lot of consensus along the way.

Clearly, "translate all of IPv4 Internet to the IPv6 Internet in both 
directions" does not work. We were not able to get very specific in 
terms of prioritizing what types of networks should be able to be 
connected to others via translators either. What we ended up with was a 
compromise between the two, with some emphasis on being able to reach 
the IPv4 or IPv6 Internet.

- Mark

Ed Jankiewicz wrote:
> I think what we are trying to get at with "my network" or "a specific 
> network" is some handy handle for the domain with address family 
> restriction that has a vested interest in translation and 
> coexistence.  While Fred is right that it is generically "an IPv4 
> network" with "an IPv6 network", the differences in the 4 sub-cases 
> MAY result in different solutions:
>
> 1.  if "my" network is the one with the minority address family, "I" 
> would/could/should be interested in bearing the burden of 
> interoperation, hence a solution that costs "me" in time and money is 
> acceptable.  As already discussed, this burden shifts over time from 
> early adopters to legacy maintainers.  The burden is either doing it 
> yourself or paying your ISP to provide the service and the addresses 
> required.
> 2.  if I am the "called" party (i.e. a server) I have a vested 
> interest in making "my" network reachable by all "calling" parties, 
> hence a distinction between IPv4->IPv6 versus IPv6->IPv4.
> 3.  more generalized solutions would be nice, but it is easy to fail 
> trying to do everything (NAT-PT) or build something that the motivated 
> party doesn't need (e.g. someone with lots of IPv4 sensors, etc. may 
> be motivated to make them reachable from IPv6, but doesn't need them 
> to "call" or be clients to anything on the IPv6 side). 4.  premature 
> optimization is counterproductive - several solutions that each solve 
> part of a big problem well can be combined and simplified, but one big 
> solution that does everything not so well (or nothing well) tends to 
> get worse in the patching.
>
> I'd say don't get hung up on having a one-word handle to hang on what 
> we mean by "the network that is constrained to one address family when 
> the rest of the world prefers the other" but I'm sure that can be 
> stated in a precise and concise manner in the charter, the scenarios 
> draft, and for introductory context for all the solutions.  "Perfect" 
> nomenclature is not necessary, but a consistent use of whatever 
> nomenclature is.
> If I had to pick one word it would be "constrained" as in "an 
> IPv4-constrained network" meaning (for whatever reason) that the 
> network interface to the rest of the world is constrained to sending 
> and receiving ONLY IPv4 addressed packets.
>
> Dave Thaler wrote:
>> The "specific" network is the one deploying the box that performs
>> translation.  The term "MY" was used at the interim, although I
>> agree with Brian that "a specific" is more readable, and is also
>> consistent with the text that follows the "MY" sentence which already
>> uses "a specific".
>>
>> -Dave
>>
>>  
>>> -----Original Message-----
>>> From: Fred Baker [mailto:[e-mail]]
>>> Sent: Monday, October 20, 2008 3:43 PM
>>> To: Dave Thaler
>>> Cc: Marc Blanchet; Brian E Carpenter; [e-mail];
>>> 46Translation; Behave WG
>>> Subject: Re: [BEHAVE] [46translation] [v4v6interim] "My" [Re: Proposal
>>> for new BEHAVE charter]
>>>
>>> Not even sure why it needs to be "a specific". We're going to ahve
>>> cross-domain issues, and if we don't think about them we will amke
>>> mistakes that will bite us later.
>>>
>>> For my money, we are connecting "an IPv4 network" with "an IPv6
>>> network", and either might be public or private.
>>>
>>> On Oct 21, 2008, at 6:33 AM, Dave Thaler wrote:
>>>
>>>    
>>>> Agree with Brian: s/MY/A specific/
>>>>
>>>> -Dave
>>>>
>>>>      
>>>>> -----Original Message-----
>>>>> From: [e-mail] [mailto:46translation-
>>>>> [e-mail]] On Behalf Of Marc Blanchet
>>>>> Sent: Monday, October 20, 2008 2:43 PM
>>>>> To: Brian E Carpenter
>>>>> Cc: [e-mail]; 46Translation; Behave WG
>>>>> Subject: Re: [46translation] [v4v6interim] "My" [Re: [BEHAVE]
>>>>> Proposal
>>>>> for new BEHAVE charter]
>>>>>
>>>>> Brian E Carpenter a écrit :
>>>>>        
>>>>>> On 2008-10-21 03:52, Magnus Westerlund wrote:
>>>>>> ...
>>>>>>          
>>>>>>>>> 1. MY IPv6 to IPv4 Internet, i.e. perform translation between
>>>>>>>>> IPv4
>>>>>>>>>                 
>>>>> and
>>>>>        
>>>>>>>> "MY" should be expanded or explained before it is first used.
>>>>>>>>               
>>>>>>> Okay, do you think the following is good enough?
>>>>>>>
>>>>>>> "My" IPv4 or IPv6 network in the descriptions below refer to a
>>>>>>>             
>>>>> network
>>>>>        
>>>>>>> with a clearly identifiable administrative domain (e.g., an
>>>>>>>             
>>>>> enterprise
>>>>>        
>>>>>>> campus network, a mobile operator's cellular network, a
>>>>>>>             
>>> residential
>>>    
>>>>>>> subscriber network, etc.).
>>>>>>>             
>>>>>> I find the use of "My" frankly a bit silly in a charter.
>>>>>>           
>>>>> I had the exact same concern but was shy to say it. now that someone
>>>>> did
>>>>> it, I support 100%.
>>>>>
>>>>> Marc.
>>>>>
>>>>>
>>>>>        
>>>>>> My Computer is cluttered with silly things like My Network Places,
>>>>>> My Documents and My Pictures, and it's all a bit annoying...
>>>>>>
>>>>>> s/My/A specific/ ?
>>>>>>
>>>>>>   Brian
>>>>>>
>>>>>> _______________________________________________
>>>>>> v4v6interim mailing list
>>>>>> [e-mail]
>>>>>> https://www.ietf.org/mailman/listinfo/v4v6interim
>>>>>>           
>>>>> -- 
>>>>> =========
>>>>> IPv6 book: Migrating to IPv6, Wiley, 2006. http://www.ipv6book.ca
>>>>> Stun/Turn server: http://numb.viagenie.ca
>>>>>
>>>>> _______________________________________________
>>>>> 46translation mailing list
>>>>> [e-mail]
>>>>> https://www.employees.org/mailman/listinfo/46translation
>>>>>         
>>>> _______________________________________________
>>>> Behave mailing list
>>>> [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/behave
>>>>       
>>
>> _______________________________________________
>> 46translation mailing list
>> [e-mail]
>> https://www.employees.org/mailman/listinfo/46translation
>>
>>   
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> 46translation mailing list
> [e-mail]
> https://www.employees.org/mailman/listinfo/46translation
>   

_______________________________________________
v4v6interim mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/v4v6interim