No, indeed, you don't really need pd_exclude, but it saves admin on not having to use separate ia_pd and ia_na
Regs Carl From: Roberta Maglione (robmgl) [mailto:[email protected]] Sent: dinsdag 19 november 2013 16:17 To: Wuyts Carl; Athanasios Douitsis; Bernie Volz (volz) Cc: [email protected]; Michael Richardson; [email protected] WG; [email protected] Subject: RE: [dhcwg] [homenet] PPP, DHCPv6 and Prefix Delegation Hello Carl, I agree with you that using DHCPv6 to number the WAN is a more common approach. In such case you don't really need PD exclude; you just need a single IPv6 address for that link and RFC 6911 in section 3.1 already defines the Framed-IPv6-Address Radius attribute to be used for this purpose. "3.1. Framed-IPv6-Address The Framed-IPv6-Address Attribute indicates an IPv6 address that is assigned to the NAS-facing interface of the RG/host." Thanks Roberta From: Wuyts Carl [mailto:[email protected]] Sent: Tuesday, November 19, 2013 7:43 AM To: Athanasios Douitsis; Bernie Volz (volz) Cc: [email protected]<mailto:[email protected]>; Michael Richardson; Roberta Maglione (robmgl); [email protected]<mailto:[email protected]> WG; [email protected]<mailto:[email protected]> Subject: RE: [dhcwg] [homenet] PPP, DHCPv6 and Prefix Delegation It's probably just a remark/side note, but pd_exclude could also be used with DHCPv6 iso RA on the WAN-link. I've not bumped in to many customers using RA on WAN links to number them, not with separate prefix nor with excluded prefix, so the typical use case will be to get/use an excluded prefix, and to assign an IP from it to the WAN link through ia_na. Regs Carl From: dhcwg [mailto:[email protected]] On Behalf Of Athanasios Douitsis Sent: dinsdag 19 november 2013 13:40 To: Bernie Volz (volz) Cc: [email protected]<mailto:[email protected]>; Michael Richardson; Roberta Maglione (robmgl); [email protected]<mailto:[email protected]> WG; [email protected]<mailto:[email protected]> Subject: Re: [dhcwg] [homenet] PPP, DHCPv6 and Prefix Delegation Hello (thanks for the answer), The uplink connection between the delegating and the requesting router will be in many cases enumerated with a prefix dictated by the Framed-IPv6-Prefix value. If this uplink prefix is going to be a part of the greater prefix that will be delegated, we would in effect have to include the value of the Framed-IPv6-Prefix in the OPTION_PD_EXCLUDE. Example, if a delegating router makes a RADIUS request and gets the following attributes in the reply: Framed-IPv6-Prefix='2001:dead:beef::/64' Delegated-IPv6-Prefix='2001:dead:beef::/56' Then the delegating router should 1)send an RA in the client uplink interface with 2001:dead:beef::/64. The uplink is enumerated with that /64. 2)Afterwards, when requested for PD, it should reply with the 2001:dead:beef::/56 to the requesting router, but excluding the 2001:dead:beef::/64 from that /56 by putting it in the OPTION_PD_EXCLUDE. So in effect, the Framed-IPv6-Prefix has been copied in the OPTION_PD_EXCLUDE option. If I have misunderstood something in the RFC or the discussion, I'd be grateful if you would correct me. Thanks very much, Athanasios On Tue, Nov 19, 2013 at 2:07 PM, Bernie Volz (volz) <[email protected]<mailto:[email protected]>> wrote: Why would it ever be copied into that option? That makes no sense to me. - Bernie (from iPad) On Nov 19, 2013, at 6:16 AM, "Athanasios Douitsis" <[email protected]<mailto:[email protected]>> wrote: (i.e. have a configuration option to use the Framed-IPv6-Prefix value in the prefix exclude option instead of an RA) Correction, the above is incorrect, as has been rightly pointed. Are there any cases where the Framed-IPv6-Prefix value will not be copied as-is in the OPTION_PD_EXCLUDE value? _______________________________________________ dhcwg mailing list [email protected]<mailto:[email protected]> https://www.ietf.org/mailman/listinfo/dhcwg -- Athanasios Douitsis
_______________________________________________ homenet mailing list [email protected] https://www.ietf.org/mailman/listinfo/homenet
