I see the "customer id" is local to the gw. This is why I don't see a new
draft is needed. However, the major challenge is even fewer bits are
available for the CPE.


On 11/8/10 9:33 AM, "Ole Troan" <[email protected]> wrote:

>> The obvious answer is to use multiple 6rd domains and set IPv4MaskLen >
>>0
>> to save bits. Then, run the dhcp server on the GW (i.e. 6rd-CE) to
>> delegate prefixes to the hosts. However, what is new in this draft
>>rfc5969
>> can't do?
>
>nothing as such. 
>the difference with gi-6rd is that you also need to encode the 'customer
>id' in the address.
>
>cheers,
>Ole
>
>>> Tina,
>>> 
>>>>>> Consider an operator facing a high subscriber growth rate.  As a
>>>>>>  result of this growth rate, the operator faces pressure on its
>>>>>>stock
>>>>>>  of available public IPv4 addresses.  For this reason, the operator
>>>>>> is
>>>>>>  motivated to offer IPv6 access as quickly as possible.
>>>>>> 
>>>>>>  The backbone network will be the first part of the operator's
>>>>>> network
>>>>>>  to support IPv6.  The metro network is not so easily upgraded to
>>>>>>  support IPv6 since many devices need to be modified and there may
>>>>>>be
>>>>>>  some impact to existing services.  Thus any means of providing IPv6
>>>>>>  access has to minimize the changes required to devices in the metro
>>>>>>  network.
>>>>> 
>>>>> what is it that makes existing softwires mesh solutions unsuitable
>>>>>for
>>>>> crossing the IPv4 only metro network? I'm thinking specifically on
>>>>>6PE
>>>>> or BGP tunnels.
>>>> Ole, thank you for asking. I didn't say that. IMHO, 6PE or BGP are
>>>>more
>>>> appropriately used in core network. There are fewer uses of MPLS in
>>>> metro network. BGP tunnels are also more appropriate for core network.
>>>> 
>>>> Essentially the authors are talking about a situation where 6rd is an
>>>> applicable technology, but needs modification to meet the constraints
>>>>of
>>>> maximized savings of IPv4 addresses and no provider access to customer
>>>> site equipment. The IPv4 address savings occur only if the customer
>>>>site
>>>> is IPv6-only, but the proposal still works for dual-stack customer
>>>> sites. The technical solution is to move the IPv6 in IPv4 tunnel
>>>> endpoint to the provider edge rather than have it in equipment on the
>>>> customer site, and use the provider gateway IPv4 address in the IPv6
>>>> prefix given to the customer site. As Ralph pointed out, the protocol
>>>>in
>>>> RFC 5969 should be able to be applied without change.
>>>> 
>>>> The authors will restructure the draft to illustrate this latter
>>>>point.
>>>> As Ralph suggested, it will be submitted with the intention of
>>>>becoming
>>>> an Informational describing an alternative deployment of 6rd to meet
>>>>the
>>>> constraints we described above. I hope this plan meets with the
>>>>chairs'
>>>> approval.
>>> 
>>> 1) is BGP tunneling not applicable because these devices don't support
>>> BGP? or not support MPLS?
>>> 2) my concern with using 6rd for this purpose is that while automatic
>>> tunneling by encoding the tunnel end point in the payload address is
>>> convenient. offering a 'sensible' sized IPv6 prefix to end users is
>>>going
>>> to be problematic. note this is purely a concern from a deployment
>>> perspective. I have no doubt that you can _use_ 6rd this way. just
>>> because we _can_, _should_ we?
>>> can you give some examples of how you deliver e.g. /56 or /64 to end
>>> users, using gi-6rd? (without expecting the SP to have a /16 of v6
>>>space
>>> obviously).
>>> 
>>> cheers,
>>> Ole
>>> 
>>> _______________________________________________
>>> Softwires mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/softwires
>> 
>

_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to