Re: enable DNS

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sat, Jun 12, 2004 at 10:11:46PM -0500, Jeff Vian wrote:
> Kenneth Porter wrote:
> > ...
> Your approach is like saying that you don't like what your small town 
> councilman (ISPs nameserver) is doing so you go the the President of the 
> United States (root nameserver) for the correct information.

An amusing analogy, but it doesn't prove anything. If the President of the
United States had a single job function - answering questions - and we have
a question that cannot be answered by our councilman, or not within a
reasonable time frame, or not reliably, why *shouldn't* we go straight
straight to the President? (Your analogy is bogus - the President's job
isn't "answer your people's questions as quickly as you can")

*Some* sort of caching name server is necessary. Whether it is a local one
used for a small network, serving only 1 or 2 computers, or whether it is
an ISP caching name servers serving 10 000, doesn't really matter. Remember,
that there are thousands more users who don't know how to set up a caching
name server, than those that do. The root name servers should be able to
handle a few hundred thousand caching name servers asking a few questions
each day (each caching name server should *not* be querying the root name
server on every single request - 'caching', remember?). If they cannot,
they definately need to be expanded.

The distributed model of DNS isn't about restricting access (although
it could be configured so, at a great administrative cost). It is
about the information itself being distributed, and each managed set
of information remaining small (ignore .com, .org, and .net, as they
obviously do not subscribe to this... :-) ). One root name server
query - ".com" - will let me contact the proper gtld for the rest of
my queries for at least the next 24 hours. This is efficient. Having
caching name servers forward to other caching name servers does not
give extra efficiency. It theoretically reduces load under *simple*
usage patterns.

As soon as somebody starts using DNSBL, or other such applications,
the "*simple*" usage patterns begin to make the load *higher*, not
*lower*. There is a line somewhere in there.

The real argument in this thread, that I can see, is whether or not
people should be using caching name servers on their home networks.
Most home networks should not bother. Why administer yet another
server that you don't have to? Why forward packets through multiple
servers? Send DNS requests straight to your ISP's domain server.

Anybody who does have a need for their own caching name server,
doesn't really benefit from forwarding requests to their ISP's caching
name server. (We're talking thousands of people who do, compared to
millions who do not)

mark

-- 
mark@xxxxxxxxx/markm@xxxxxx/markm@xxxxxxxxxxxxxxxxxx __________________________
.  .  _  ._  . .   .__    .  . ._. .__ .   . . .__  | Neighbourhood Coder
|\/| |_| |_| |/    |_     |\/|  |  |_  |   |/  |_   | 
|  | | | | \ | \   |__ .  |  | .|. |__ |__ | \ |__  | Ottawa, Ontario, Canada

  One ring to rule them all, one ring to find them, one ring to bring them all
                       and in the darkness bind them...

                           http://mark.mielke.cc/



[Index of Archives]     [Current Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [Yosemite Photos]     [KDE Users]     [Fedora Tools]     [Fedora Docs]

  Powered by Linux