> There my view on the issue is quite clear: For a simple node (phone, or a PDA) or for a node in a simple network (PC in home environment, etc. - no real administration), there has to be a way to use the DNS without running _any_ additional protocols to additional servers. I believe that the current 'dns discovery' draft fills that whole rather well. (Though, the title of the document does not really reflect the content - to my opinion).
I absolutely agree. In other word, there are typically three players, 1) end-host 2) site network 3) ISP network and "Stateless DNS discoery" is a zero-configuration method mainly for 1). When the administrator of 2) wants to prepare DNS servers in his/her site with assigning the well-know-site-local-uni-cast-addresses to them, 1) simply queries to them. When the ISP prepares DNS servers in its backbone, the CPE router of 2) acts as a dual-sited DNS proxy to relay queries to the the well-know-site-local-uni-cast-addresses of the ISP's site, or to the global addresses which are informed via ISP-to-Customer (or PE-to-CPE) configuration mechanism such as DHCPv6, SLP or papers. Toshi Yamasaki / NTT Communications -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------