Hacker News new | past | comments | ask | show | jobs | submit login

This is surprisingly hard to explain off the cuff. In DNS, zones are separated by dots, and a fully-qualified domain name ends in a dot. The zones exist because there are too many domains for just one server to handle and this way zones can be distributed and managed separately. DNS queries right to left, for “Google.com.” it would look at “.” and query the root servers asking if they know “Google.com.”, they could reply saying ask “.com.” at this IP. Then the .com. servers look up Google.com. and (eventually, there’s one more lookup for a Google server) return its IP address. Google.com. loads in your browser because an A record is returned by a Google name server. “ai.” is an A record returned by the name server for .ai, as registered in the root DNS servers. Similarly, Verisign could register an A record for .com. But they haven’t... um... Google for how DNS works for a more thorough and perhaps animated example.



> The zones exist because there are too many domains for just one server to handle and this way zones can be distributed and managed separately.

Though that's not really true in practice. Com hasn't exploded under the amount of traffic it faces, and merging in the other zones would only be about twice the load.


Com only handles the NS records and glue records for its direct descendants though. If there were no zones below com, its nameservers would melt from the update requests alone.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: