Thanks again for the comments Jon. All noted and I agree with Mark's response.
Tim On 6 Mar 2013, at 04:34, Mark Townsley <[email protected]> wrote: > > Replying, changing the very misleading subject. > > Thanks for the comments, please see inline. > > On Mar 1, 2013, at 2:45 AM, Brzozowski, John wrote: > >> Tim, >> >> >> Per your request below here are some comments pertaining to the section >> you mention below. >> >> * The section seems to hint at two extremes stable on one end and very >> dynamic i.e. at each CER reboot. I think it is important to note that >> there is middle ground here and that providers typically have good reason >> for requiring prefix changes, largely related to network capacity >> management. It is not clear to me if this need to be called out in the >> text, however, I thought you should know. >> * Agreed regarding /64 subnet lengths >> * This section does not differentiate between a residential service and a >> commercial or SOHO service. Is this outlined elsewhere in the draft? >> Does it need to be? > > That's more of a charter question. While we should aim to create technology > that is generally useful outside the home, the defined scope is the home. > Among other things, this should help underscore the auto-configuration > requirement as the further you dip into SOHO territory, the more likelihood > that there is an administrator of some sort (whether it be a remote worker's > IT department or otherwise) becomes more and more prevalent. > > While the "SO" absolutely must be able to work with our "HO" (what looks to > us like a special type of additional ISP, if you will), we stop short of, > say, trying to define the VPN router's security policy or whether it should > use IPsec Tunnel mode or GRE with Transport Mode or SSL or... > > (Personally, I would like to see homenet's technology have impact outside the > home - we're tackling some important problems IPv6 has had for a while.) > >> * Note - some devices today can only utilize a /64 and nothing more these >> would not break. > > Good point, and this brings up a general issue of working with existing > non-homenet routers that support IPv6. The document should probably highlight > this. > >> * No NPTv6 please > > Section 2.4 > > "As such, neither IPv6 NAT or NPTv6 is recommended for use in the homenet > architecture." > >> * ULAs support could be introduced at a later date > > Why wait? > >> * Agreed regarding bridging avoidance >> * How is "relatively stable prefix" defined? > > Relative to changing the prefix whenever the CER is reset: > > "Some ISPs may offer relatively stable prefixes, while others may change the > prefix whenever the CER is reset." > > Feel free to offer up a better description. I think the salient point ends up > being that the home network is going to have to be resilient to prefix > change. > > >> * Good text regarding PIA >> * Regarding renumbering also note that operators take precautions to >> ensure this is seamless as well including lease time lowering in advance >> of the renumbering event. We have been doing this for years. > > Consistent with RFC 4912 or not? (because that is what we are pointing to > right now) > >> * Forced renumbering for privacy sounds unfortunate but based on various >> emails I have seen lately this seems to be something that must be >> supported. > > Yes, sadly. > >> * Agreed that service discovery is essential >> >> I hope these comments are useful, I recognize I was not overly verbose. I >> hope this is a feature not a bug. Please let know if you wish to discuss >> further. > > Thank you for the review and input, John. > > - Mark > >> >> John >> -----Original Message----- >> From: Tim Chown <[email protected]> >> Date: Wednesday, February 27, 2013 5:56 AM >> To: "[email protected]" <[email protected]> >> Subject: Re: [v6ops] new draft: draft-shishio-v6ops-dpvt >> >>> >>> On 26 Feb 2013, at 14:07, "Brzozowski, John" >>> <[email protected]> wrote: >>> >>> >>> [jjmb] incorrect I have both, just stating that I have both. Apologies if >>> I was not clear. I also have a good # of home routers actively using IPv6 >>> enabled broadband today, ~3%. >>> >>> >>> >>> >>> Congratulations, btw. That must be a large number when converted to real >>> customers :) >>> >>> >>> The homenet architecture assumes a CPE rather than a single host. It >>> discusses ISP allocations in section 3.4.1 of >>> http://tools.ietf.org/html/draft-ietf-homenet-arch-07. As that text >>> is in WGLC, any comments on that section would be welcome (preferably on >>> the homenet list). >>> >>> >>> Tim >> >> _______________________________________________ >> v6ops mailing list >> [email protected] >> https://www.ietf.org/mailman/listinfo/v6ops >> _______________________________________________ >> homenet mailing list >> [email protected] >> https://www.ietf.org/mailman/listinfo/homenet > > _______________________________________________ > homenet mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/homenet
_______________________________________________ homenet mailing list [email protected] https://www.ietf.org/mailman/listinfo/homenet
