Root server cannot be a forwarder?

April xiaoxia2005a at yahoo.com
Sat Oct 21 04:21:46 UTC 2006


April wrote:
> Barry Margolin wrote:
> > In article <ehbkg4$1g42$1 at sf1.isc.org>,
> >  Kevin Darcy <kcd at daimlerchrysler.com> wrote:
> >
> > > Stephane Bortzmeyer wrote:
> > > > On Fri, Oct 20, 2006 at 04:43:55AM -0700,
> > > >  April <xiaoxia2005a at yahoo.com> wrote
> > > >  a message of 18 lines which said:
> > > >
> > > >
> > > >> I'm not talking about the Roots for the Internet, but a namespace in
> > > >> general.
> > > >>
> > > >
> > > > I do not recognize DNS vocabulary and ontology. On a BIND mailing
> > > > list, using the DNS terminology seems a good start :-)
> > > >
> > > I don't see any problem with the terminology: a root server is a server
> > > that answers authoritatively when queried for the root zone. Whether the
> > > client or the server happens to be on the Internet shouldn't have any
> > > bearing on the terminology used, it's only the context of the transaction.
> >
> > But that's not what people were referring to by "root server" in this
> > thread, either.  They're using the term "root server" to mean "server
> > for the root of my domain".  In other words, on the Acme Corporation
> > network, the server that hosts acme.com is their "root server".
> >
> > I think people really need to learn to be less pedantic.  When context
> > makes the meaning clear, we can survive sloppy terminology and
> > ambiguity.  The fact that someone asks a question like the one in the OP
> > indicates that they don't have much expertise in this area, so do you
> > really expect them to be fully familiar with our terminology?  We're
> > smart people, we can figure out what they mean even if they don't use
> > the exactly correct words.
> >
> > --
> > Barry Margolin, barmar at alum.mit.edu
> > Arlington, MA
> > *** PLEASE post questions in newsgroups, not directly to me ***
> > *** PLEASE don't copy me on replies, I'll read them in the group ***
>
>
> Just found something you posted today I think,
>
> "> Referring who to the parent nameservers?  Clients almost always have
>
> > "stub resolvers", which do not implement iteration by themselves.  They
> > send queries with the Recursion Desired flag set, and depend on the
> > server to perform recursion to look up remote names.
>
> where is this "remote names" coming from?  Is this a DNS terminology?
> That is called something like "names the server is not authoritative or
> cached" if you want to speak DNS, right?  ;-)


Now let's look at the DNS in this post:

>> Referring who to the parent nameservers?  Clients almost always have
> > "stub resolvers", which do not implement iteration by themselves.  They
> > send queries with the Recursion Desired flag set, and depend on the
> > server to perform recursion to look up remote names.

The stub resolver also does not do recursion either, though it sends
out recursive queries.  Both iteration and recursion are nameserver
resolution processes, which are not parts of a stub resolver's
business.   The stub resolver sends recursive queries with Recursion
Desired flag set, and the nameserver will look at the authoritative
data and cache first, if not found, then start recursion to look up the
"remote names", which in "DNS" means the names that not authoritative
by the server or not currently in the servers cache.



More information about the bind-users mailing list