> Iljitsch van Beijnum, October 12, 2007 10:45 AM > > On 12-okt-2007, at 14:23, Eric Voit (evoit) wrote: > > > Many DSL providers already use DHCP with Option 82 (added at the > > DSLAM) > > for location information. The EAP information transported from the > > client would be used incrementally for subscriber (or CPE) identity. > > Yes, I can see how a DHCP solution is attractive from this > perspective. > > > (As an FYI for triple play, it is important that "trusted" customer > > specific CPEs can be identified.) > > Are we then talking about a protocol that is only exchanged > between an ISP-supplied CPE and the DSLAM? In that case they > can use any protocol that they want because there is no > interaction with the end- user or any equipment that's not > under the control of the ISP.
I don't believe this is the case. SPs still have a UNI demarcation between the home and the network. SPs desire a standardized protocol which matches to their existing operational experience here. Beyond that, off-the-shelf CPE is still relevant. In fact this email will traverse my Verizon FiOS triple play router. It turns out this is a typical off-the-shelf Broadband Router, except with some tweaks such as the hiding of the WAN PPPoE credentials & configuration. However before I got my triple play Router, I just had FiOS broadband service. In that case I used an off-the-shelf Linksys router, and I was able to enter my PPPoE credentials myself. In other words, the CPE to Network protocol stayed the same across my Verizon FTTH phys as the service mix and even CPE changed. Standardizing such parts is *very* valuable for Operations & troubleshooting. > However, from the list of requirements it seems that the DSL > Forum is looking for something that goes from the end-user PC > to the DSLAM or from a third party gateway installed and > operated by the end-user to the DSLAM so the protocol needs > to make sense both on commercially available home gateways > and on regular multi-purpose operating systems. Sequencing is always an issue since it impacts the phased emergence of deployment models. PPPoE was PC client software long before it made it into Broadband routers. > >> Another good way to meet the DSL Forum's requirement would > be to run > >> PANA over IPv6 link-local addresses. > > > What you describe is certainly possible. But to me it seems more > > complicated for failure mode debugging than the DHCP Auth proposals. > > (And remember, DSL providers already use DHCP & Option 82.) > > So how would a DHCP solution work when the user connects: > > - a Windows XP or MacOS X machine but Microsoft/Apple won't > be including the new DHCP option for another 12 months > - a Windows 98 machine that is no longer getting updates from > Microsoft > - a no brand home gateway which can't be upgraded by the user > - a machine that runs DHCP for IPv4 and DHCPv6 for IPv6 > - a machine that runs DHCP for IPv6 and runs IPv6 but not DHCPv6 > - a machine that doesn't run IPv4 but IPv6 + DHCPv6 > - a machine that doesn't run IPv4 but IPv6 without DHCPv6 All good questions, it will be interesting to see which of these are determined to be in-scope if the DSLF request for a WG draft in this space is accepted :-) Eric _______________________________________________ Int-area mailing list [email protected] https://www1.ietf.org/mailman/listinfo/int-area
