We now know how DNS works without discussing all the nasty things like protocols and how to deal with the software end. If you’re a sys admin, it’s crucial that you know these things.
The process begins with the client submitting a DNS query to a DNS server. This is because client software, such as Web browsers, are not designed to hunt down IP addresses for themselves. If they did, the Internet would be full of DNS queries.
Most DNS servers belong to ISPs. They hold little information about a
domain, but simply look up information for you and I, the client computers.
These machines are no different from many machines we use hardware wise,
they simply have special Software that does most of the work.
A client can get an IP address two different ways. The first way, and the fastest, is called iterative resolution. Iterative resolution works in a fairly straightforward manner. Clients simply ask the server to resolve a domain name for them. The server accesses its local database, finds the IP address, and sends it back. If the server doesn’t find the address, it sends back a “DNS not found” error.
The second DNS resolution method is a little more complicated. The client requests a name resolved by the server. The server will first do a normal iterative look up, checking its own local database.
The next step comes in when the name isn’t found in the DNS server’s local database and is explained best through example:
Using the made-up name computer.lan.wan.com, the DNS server will first query the root domain server for the .com portion for the IP address of the wan.com name server. Then, the DNS server will query that name server for the name server of lan.wan.com. Next, the DNS server will query that name server for the IP address of the computer name “computer”. The DNS server will finally return the IP address of computer.lan.wan.com to the client, and then save the IP in its database for future reference.