ietf
[Top] [All Lists]

Re: WebAddress resolutions

2001-09-21 09:20:03

In message <003901c142b2$63e04aa0$0100000a(_at_)Thor>, Thor Harald Johansen 
typed:
If I'm not completely mistaken, this is good old fashion bull. It's a joke.
 
correct - i guess i should have put a
:-)
for the clueless

note that given most names are of the form
foo.com
and are issued by a small number of authorities rather than the
original design goal of DNS being very very
distributed (and replicated) and hierarchical, actually 
a fixed allocation of client/root server, and a hash lookup
implementation in the root/gTLD servers would actually work for about
95% of systems and maybe a lot better than currently:

given typical findings on the effectiveness of the access protocols
and of cacheing, e.g. see
"DNS Performance and the Effectiveness of Caching," 
by Jaeyeon Jung, Emil Sit, Hari Balakrishnan, and Robert Morris, is at
http://nms.lcs.mit.edu/papers/dns-imw2001.html

it might be wise to move to a more centralised system and be
honest:-)



Domain names (www.something.xyz) are resolved using a network of DNS (Domain
Name Service) servers. There are in fact a few central servers to this
system, but they usually don't serve the crowds. This is the job of each
ISPs own DNS servers, wich are updated frequently with new entries from the
central DNSes.

These servers also contains information for reverse resolving. That is to
find out wich IPs have been pointed to by a domain. The Internet would work
fine without domain names. It would just be a bit harder to use (you'd have
to type http://64.12.50.249/ instead of http://www.cnn.com/). I belive DNS
came a long time after IP addresses. :)
 
and took nearly as long to depoy as ipv6 and ip multicast:-)

i recall doing a comparison of druid and bind at UCL in the mid-late
1980s....about 7 years after we switched from ncp to tcp/ip...

 cheers

   jon



<Prev in Thread] Current Thread [Next in Thread>