Le 2012-04-10 à 12:07, Wojciech Dec a écrit :

> The pre-standard-still-being-defined DHCPv6 option was not tested, and CLI 
> based configuration including scripting was applied.
> 

> Question closed.

For you, that's your choice, fair enough.

But not closed for the WG: members deserve to know how CEs are proposed to know 
whether they must work in mesh or hub-an-spoke topology (even if MAP remains 
experimental).

RD


> -Woj.
> 
> On 10 April 2012 12:01, Rémi Després <[email protected]> wrote:
> Wojciech,
> 
> This isn't answers to questions I asked.
> They remain open.
> 
> RD
> 
> Le 2012-04-10 à 11:51, Wojciech Dec a écrit :
> 
>> Remi,
>> 
>> you're apparently confusing matters. There is no need to have a DHCPv6 
>> option, or a million node deployment to test MAP implementations. DS-lite is 
>> a good example, with implementations and standards track before the DHCPv6 
>> option.
>> 
>> Needless to say, if you're implying that tests of MAP without testing the 
>> standards based DHCPv6 option are insufficient, then any test of 4rd-u or 
>> anything for that matter without using the fully standards DHCP option would 
>> be equally flawed. 
>> At the very least however, MAP does not need to prove on thing: 
>> Compatibility with IPv6, which 4rd-u would need to.
>> 
>> -Woj.
>> 
>> On 10 April 2012 11:37, Rémi Després <[email protected]> wrote:
>> Hello, all,
>> 
>> We have heard many times that MAP is completely specified, and has been 
>> extensively tested.
>> Yet:
>> - mapping rules of tested configurations have not been provided
>> - several missing points of the MAP-T+E specification have been identified 
>> (ref (*) www.ietf.org/mail-archive/web/softwires/current/msg04049.html).
>> 
>> This mail is just about ONE of these, the hub-and-spoke issue.
>> It has been discussed several times but AFAIK still without a complete 
>> answer.
>> 
>> The difficulty is that:
>> - The MAP-DHCPv6 draft has no parameter to indicate whether the ISP-chosen 
>> topology is mesh or hub-and-spoke.
>> - According to the MAP-address-and-port draft, "each MAP node in the domain 
>> has the same set of rules".
>> - As answered in the mail below, the choice "needs to be provisioned. either 
>> explicitly or implicitly (via the rules)".
>> 
>> Questions I have are then:
>> - Is the choice provisioned explicitly, implicitly, or possibly both?
>> - How?
>> - Which tests have confirmed that it worked?
>> 
>> Answer by any one who asserts he or she understands how MAP works will be 
>> welcome.
>> 
>> Thanks,
>> RD
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> > De : Ole Trøan <[email protected]>
>> > Date : 2012-03-14 14:29
>> > À : Rémi Després <[email protected]>
>> > Cc : Tomasz Mrugalski <[email protected]>, Softwires WG 
>> > <[email protected]>
>> > Objet : Rép : [Softwires] Question about hub-and-spoke operation in MAP
>> >
>> > Remi,
>> >
>> >> I couldn't figure out by how CEs can be required to work hub-and-spoke 
>> >> without some DHCPv6 indication:
>> >> - If two CEs apply the same BMR to their delegated IPv6 prefixes, how do 
>> >> they know whether their ISP expects direct paths between them (mesh) or 
>> >> BR hairpinning (hub-and-spoke)?
>> >>
>> >
>> > that's correct it needs to be provisioned. either explicitly or implicitly 
>> > (via the rules).
>> >
>> > 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