Section 2 of TR-101 says: "The principle guiding this specification work is that the resulting functions should enable an as smooth as possible migration process from an ATM based aggregation network to an Ethernet based aggregation network."
Terminal originated PPPoE is mentioned also in TR-101 (e.g., Figure 15). >From customer's point of view, it would be natural to think that a *smooth* migration path includes support for existing PPPoE terminals to be able to migrate to Ethernet based aggregation network without necessarily upgrading their legacy CPEs to RGs. In this thread, when the discussion comes to end-host support, I always see inconsistency between what is stated in the DSLF requirements (and TR-101) and what is stated by proponents of DHCP authentication. If IPAuth-8 and IPAuth-9 are not really requirements, why not re-send revised requirements officially from DSLF to IETF? Unless requirements are revised, I don't think we can make a progress. Yoshihiro Ohba On Thu, Nov 22, 2007 at 01:38:23PM +1000, Richard Pruss wrote: > Alper Yegin wrote, around 22/11/07 6:43 AM: > >>Alper Yegin wrote, around 21/11/07 5:29 PM: > >>>And how about on the PCs? > >>Have you read the requirements? > > > >Yeah I did. Have you? > > > > > >IPAuth-8 Must handle L3 CPE device authentication and end-device (PC) > >user based authentication (likely with L2 CPEs in the latter case) > > > >IPAuth-9 Should be simple to implement on client (PC or CPE) > > > > > > Ahh okay. I understand how you have got confused. It is a little > subtle. The key thing is: > IPAuth-6 Must fit into TR-101 operational model > And then we always send TR-101 along. > > TR-101 has two models > > - one is the bridged and the only identity is done by port marking DHCP > in Access Node. > > - the other is routed with PPPoE running from the RG, PPPoE tag > insertion on the access node marks the site, and obviously PPP does the > identity. > > If you really want to dig you will find the requirements in device > originated PPPoE in TR-68 which is the document on RG's. > > So in short TR-101 does not have username and password across what is > called the T-interface in the DSLForum i.e. the customers lan, so we do > not need to consider a migration path for them. > > The main change in WT-148 is to move to an environment where one has > multiple services points on the ethernet as opposed to TR-101 which all > services came through a single BRAS. > > Regards, > Ric > > > > > > _______________________________________________ > Int-area mailing list > [email protected] > https://www1.ietf.org/mailman/listinfo/int-area > _______________________________________________ Int-area mailing list [email protected] https://www1.ietf.org/mailman/listinfo/int-area
