"to deliver the prefix length via DHCPv6, at which point you may as well also deliver a default gateway, and complete the circle that starts the deprecation of RA."
Agree with the former, not so much the latter. Regardless of the presence of DHCPv6 infra., not all hosts can / want to do it - thus RAs still needed (and therefore RSs still desirable). Also - talking about link locals being invalid in this ad-hoc situation is invalid ... There are other (better) ways to do name resolution in that type of situation, even assuming DNS is an option (which is a stretch itself). /TJ PS - if anyone is counting, my vote is to do DHCPv6 only when the RA says to(M|O), or in the absence of RAs ... I think that is the "most v6'ish" way to do it while also covering the below-mentioned ad-hoc situations. Sent from my Verizon Wireless BlackBerry -----Original Message----- From: "David W. Hankins" <[EMAIL PROTECTED]> Date: Mon, 13 Oct 2008 09:49:40 To: DHC WG<[EMAIL PROTECTED]> Cc: IPV6 List Mailing<ipv6@ietf.org> Subject: Re: [dhcwg] Brokenness of specs w.r.t. client behavior with M&O bits -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 -------------------------------------------------------------------- -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------