> Stateless address autoconfig mode you're talking about is about > advertising a /64 prefix on the link, with addrconfig flag set, and > onlink flag not set --right? While the onlink flag gives no > information how the hosts could communicate between themselves, it > still gives the ability to connect multiple hosts on a link -- so it > cannot be used by the ISPs to force "1 user/prefix" -model, right?
Not if the link to the ISP is a pt-pt link, and all the case I've seen, even if the L2 is broadcast capable such as in cable networks, the topology visible to IP seems to be pt-pt. The ISP can also do strict enforcement if they'd like on top of this. For instance, by just letting the first IPv6 address they see coming over the link from the subscriber as the only IPv6 address that can be used; packets to/from other addresses could be filtered out by the ISP. Thus in this tussle between the ISP (and its, IMHO misinformed, desire to charge per host) and the user's desire to share the bandwidth they are buying across multiple devices, it is not fruitful to try to use protocols to try to force the ISP to change their thinking; it isn't hard for them to do filters as a last resort to prevent what they think they need to prevent. What we can do from a protocol perspective (and implementation perspective) is to make it easy for the ISP to do the right thing - explicitly delegate a prefix (/48 by default as well as /64) to the customer. > > Thus I think the ISP that wants to charge differently for one IPv6 address > > per customer compared to a /64 or /48 prefix can do this as easily as it > > can in IPv4. > > I don't know how this relates to the above, but this seems obvious. The basis for needing ndproxy in this case is that the ISP doesn't want to explicitly delegate a prefix to their customer. I see two possible reasons for this behavior (and there might be more) - the ISP wants to provide tiered services with a single device service at the bottom i.e. only allow one host/IPv6 address - it is too hard for them to explicitly delegate prefixes What else might make then not do this? Erik -------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
