Re: [Softwires] Working group last call for draft-ietf-softwire-map-05

[<[e-mail]>]

Hi Ole,

I think that the change that you've proposed would help. But actually the
worst offender is in section 5:

Port-aware IPv4 entries in the Rules table are installed for all the
Forwarding Mapping Rules and an IPv4 default route to the MAP BR.


What about if this read?:

Port-aware IPv4 entries in the Rules table are installed for all the
Forwarding Mapping Rules and an IPv4 default route via a tunnel to the MAP
BR.

Cheers,
Ian



On 10/04/2013 15:20, "Ole Troan" <[e-mail]> wrote:

>Ian,
>
>> Thanks for that. So we actually aren't too far apart. There needs to be
>>a DMR (or equivalent) that has the IPv6 address of the BR.
>> 
>> But, this is one thing that's been confusing me since the publication
>>of (I think v03), where the DMR is no longer described, or mentioned
>>anywhere in the MAP base draft. However it does still exist in the MAP
>>DHCP option draft. But neither of the drafts actually describes what it
>>is and how to use it. The current map 05 draft only discusses the use of
>>an IPv4 route, which I think is misleading ­ the IPv6 address of the BR
>>is also needed (as Ole's pointed out in his mail).
>> 
>> A DMR, or something functionally equivalent is necessary, but this is
>>not just an IPv4 route. It has to have the v6 address of the BR, used to
>>build a tunnel. The v4 route points then to that tunnel.
>> 
>> So, the reason that I said it's incompatible is that the current text
>>in the MAP base draft doesn't describe any way of getting the v6 address
>>of the BR other than those that can be provided with a BMR and FMRs.
>>I.e. An IPv4 default route with an IPv4 next hop which is in a map
>>domain which the CPE has an FMR for. lw4o6, which doesn't use mapping
>>rules cannot calculate the v6 address of the concentrator from a v4 next
>>hop address.
>> 
>> I realise that this is not how it is meant to work (and you've
>>confirmed that), but this is the only way that you could make it work
>>given the text that is actually in the draft and that's what I think
>>needs to be changed. This would bring the MAP/lw4o6 & unified CPE drafts
>>back in lineŠ
>> 
>> Š then we can have the discussion about what the best DHCP mechanism to
>>provision the client with the v6 address of the concentrator is :-)
>
>please propose text.
>the MAP base document does not specify the provisioning mechanism itself.
>but it does specify which parameters need to be provisioned.
>
>section 7 states:
>
>For a given MAP domain, the BR and CE MUST be configured with the
>   following MAP elements.  The configured values for these elements are
>   identical for all CEs and BRs within a given MAP domain.
>
>   o  The Basic Mapping Rule and optionally the Forwarding Mapping
>      Rules, including the Rule IPv6 prefix, Rule IPv4 prefix, and
>      Length of EA bits
>
>   o  The IPv6 address of the MAP BR.
>
>   o  Hub and spoke mode or Mesh mode.  (If all traffic should be sent
>      to the BR, or if direct CE to CE traffic should be supported).
>
>
>section 5.4 states:
>   To reach IPv4 destinations outside of the MAP domain, traffic is sent
>   to the configured address of the MAP BR.  On the CE, the default can
>   be represented as a point to point IPv4 over IPv6 tunnel [RFC2473] to..
>
>
>would it help you if section 5.4 states the "configured IPv6 address of
>the MAP BR"?
>
>cheers,
>Ole
>
>
>
>