Slashdot Mirror

← Back to Stories (view on slashdot.org)

The Worldwide Domain Battle

Posted by timothy on Sunday March 21, 2004 @05:58AM from the translation-engines-reloaded dept.

pledibus writes "The New York Times's Sunday magazine contains an interesting article, Get Out of My Namespace, about the spate of conflicts over website names. The author synthesizes ideas from computer technology, law, history, onomastics, cultural anthropology, and probably a few other areas, and does a pretty nice job of it."

2 of 183 comments (clear)

Min score:

Reason:

Sort:

Reg Free Link by Anonymous Coward · 2004-03-21 05:59 · Score: 5, Informative

Reg Free link
Re:Good article, but has a couple of myths in it by Anonymous Coward · 2004-03-21 07:09 · Score: 5, Informative

There are many root servers, but there is only one primary database of domain data. It is in Reston. All root servers get their information from Reston.

Domains are mapped to nameservers in their domain record, not in DNS queries. This data is in the root servers (for the TLD, not for '.'), and changes do, in fact, propagate out to the other root servers when they ask the master for updates.

DNS data itself can be seen to propagate out, when you include the concept of TTL (time-to-live) for the data. You don't always query authoritative nameservers for an address -- it would overload them (and where would you stop? you'd have to go all the way up to the root servers to be sure you were getting good info). You ask your local cacheing nameserver, run by your ISP, who checks its cache to see if it already "knows" the answer, and whether the answer is "older" than its TTL. If it is older, it usually queries the authoritative nameserver for the domain. If it is younger, it just returns the same value as before.

So the data doesn't propagate per se, but the awareness of it does, and not instantly. Sometimes not even quickly.

And yes, your browser caches the response too, but that has nothing to do with DNS or TTL.