OK – another deep hole – just love these.
Simplest answer is that there is a server (arbitrarily chosen by the appropriate admins) which is THE authority for example.com, and its associated address range 168.192.in-addr.arpa (which covers all of 192.168.x.y).
note the reversing of the byte ordering for the address range.
Other servers can get to know the data by one of two (or three) basic methods.
Assume that example.com is known by the TLD (Top Level Domain) name servers (.COM in this case) as being the SOA (Start Of Authority) for example.com. If nobody has any data for example.com, then their query will go all the way up to the root name servers for the internet.
So, let’s imagine a query for www.example.com. Nobody’s name server originally has the data for that value – it’s just been created. But the root name servers DO know who’s responsible for all information about example.com – so they query it.
It comes back and says that the IP address for www.example.com is 192.168.1.1
At that point, every name server along the original query route now gets that answer, and caches it for some period of time (minutes, hours, days, it depends on how it was set up – not important for the moment).
Now when the next query comes up for example.com, there are a number of name servers that have that data cached, and can answer – but with the proviso that it’s non-authoritative – meaning that it (the answer to the query) was retrieved using cached data. It’s non-authoritative because during the lapsed time, the real answer COULD have changed. Probably not, but it could have.
So, if you’re running example.com, and it’s getting popular, you might want to have some other name servers keep copies of your data – and keep it current.
Now you’re getting into defining other servers as being authoritative for your domain – or zone. Bear in mind – a zone is an arbitrary division. It can be an entire domain, or just a chunk of it. Up to the administrators exactly how they have divided it up.
If xyz.com agrees to do so (at example.com’s request), then they can become authoritative for example.com.
Similarly, other servers (if allowed by example.com) can also carry copies of the entire zone (or domain if set up that way). Now example.com is “populating” a number of servers.
Commonly, there are caching-only servers which simply keep copies of frequently queried values (www.google.com, www.yahoo.com, www.espn.com, etc.) for fast reference, since most people don’t need authoritative data.
If this is not enough, send me a private message and we can talk by phone.
I also recommend the book DNS and BIND by Paul Albitz and Cricket Liu. They go into much more detail than I want to typing.