At 2:47 PM +0200 8/31/05, Brian E Carpenter wrote:
>That is about 1/3 of the total. It doesn't surprise me at all that
>so many bogus queries arrive - everybody who mistypes a TLD or
>misconfigures a default domain generates bogus queries, and this isn't
>going to change. The question is whether .local is a *significant*
>part of this load. The limited data I have suggest not, but I'd like
>to see publicly available data: what fraction of those NXDOMAINs are
>due to .local?
The question maybe should be whether .local will become, not is, a
significant part of this load. If you prevent mDNS, the main
proponent of the .local namespace, from becoming a standard, the
number of those names will remain low. If it becomes a standard and
implementers use that namespace more, the load will of course
increase.
I have only skimmed the mDNS proposal, but I don't think this is correct. mDNS
defines .local as a local namespace and requires that queries in this namespace
not done in the DNS. An implementation that forwarded .local queries to the DNS
would not be compliant IMO.
LLMNR, on the other hand, effectively requires this behavior if you set up
something like .local (it could just as easily be .foo), because it says
that when you have both DNS and LLMNR access you try DNS first and if
that fails fall back to LLMNR.
Ned
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf