Hi Steve, As usual we agree at the bottom line. Good point about the routers not really being src/dst deliberate usually too. I missed that when typing.
But as chair I want to say we should ship this spec with the unicast approach as Bob and others have stated. I know of 3 vendors now that are going to ship small devices late 2002 and for sure early 2003 and we need this problem fixed. OK I will admit I am one of those vendors too so I am biased. thanks and again great write up, /jim > -----Original Message----- > From: Steve Deering [mailto:[EMAIL PROTECTED]] > Sent: Wednesday, May 01, 2002 10:11 PM > To: Bound, Jim > Cc: [EMAIL PROTECTED] > Subject: RE: Stateless DNS discovery draft > > > At 7:07 PM -0400 5/1/02, Bound, Jim wrote: > >Good analysis but something about it don't sit well. My > first response is > >that Routers are intermediary nodes and require > configuration too so have > >all the properties that come with that flaw. > > Jim, > > The pedantic answer is no, routers are not necessarily intermediaries > according to my definition of that term. Intermediaries are entities > or sets of entities that a seeker sends *to* and/or receives *from*, > in order to acquire needed info about a target. If those entities are > not on the same link as the sender, yes, you need routers to enable > that sending and/or receiving, but the routers themselves are not the > destination or source of the seeker-intermediary communication (unless > a router coincidentally happens to be the home of either seeker or > intermediary). > > The more pragmatic answer is sure, routers are intermediaries. If you > are going to allow seekers and targets to be on different links, you > necessarily rely on an intermediary of some sort. The goal (for > robust plug-and-play) is simply to eliminate *unnecessary* > intermediaries, because each intermediary is a source of additional > potential failures. For an Internet of more than one link, we > obviously need routers to enable communication; the question is > whether or not impose a requirement for *more* intermediaries that > those. > > Note also that routers generally run protocols designed to maximize > fault tolerance, by allowing arbitrarily redundant topologies, by > not having single points-of-failure, and by ensuring that if > a physical > path exists from A to B, then packets can be (best-efforts) delivered > from A to B. That's generally not the case for autoconfiguration > servers. > > >...my domain required management to be set up the way I personally > >want it to be set up not the way the canned techno parts came to me > >via UPS. > > Fine, the IPv6 stateful autoconf option is there for those who want > to do that. But my neighbor *does* want it to just work out-of- > the-UPS-box, and I want IPv6 to be reliably usable by all my > neighbors, > not just the ones who are geeks. (Well, here in Silicon Valley, it's > probably the case that all my neighbors are geeks, but you know what > I mean... :-) > > Steve > > -------------------------------------------------------------------- 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] --------------------------------------------------------------------
