On Fri, Aug 20, 2010 at 10:06 AM, Christian Huitema <huit...@microsoft.com> wrote: >> yes. this seems like a case of something that looked like a great idea >> 12+ years ago (rfc2461 was published in 1998, LOTS of things have >> changed since that time) but is upon reflection maybe not a great >> idea. > > As codified 12 years ago, redirects were supposed to solve two problems: > multiple subnets on the same link; and non-broadcast networks. The discussion > on this list makes it pretty clear that the "multiple subnet" use case is > rare enough, that redirects have harmful effects, and that they should not be > enabled by default. But if we do not have something like redirects, how do > we handle non-broadcast networks?
I think my stance (and jareds and mikael and stenar) was that having redirects is good/acceptable, having them on by default is not good/acceptable. So... for 'non-broadcast' networks you enable that feature on the interface(s) in question. Again, I can imagine some helpful instances where the vendor codes the router to say: "Hey, that's a non-broadcast interface, here I fixed it for ya!" though as Jared points out making that sort of thing very, very obvious in the config AND docs would certainly be nice as well. -Chris -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
