Structure and operation of DNS

DNS is a hierarchical decentralized name correspondence system a bit like a computer directory tree structure at the top is a "root" Then it is divided into several basic category names such as comorgedu, and then the following is the organization name such as Sonytoshibaintel, and then the host name such as WWWMAILFTP. Because the internet was originally originated from the United States, so there was no country domain name, but with the subsequent development of the Internet DNS also added such as Twhkau domain name. So a full DNS name is like this www.xyz.com.tw and the entire name corresponds to an IP address.

At the beginning there were only six organization categories under root

Category name represents meaning
EDU Educational Academic Unit
ORG Organizational structure
NET Network communication Unit
COM Company Enterprise
Gov government Agencies
Mil Military Unit

But since the name of the organization has been opened, various names have come out, but in any case the rule of naming is best suited to the nature of the site. In addition to the original category information is administered by the NIC (Network information Center) of the United States, the following categories are managed by the country's NICs, respectively. This structure looks like this.

The DNS of the organizations in the structure is administered by the organization or its authority (usually when you apply for a domain domain name, you specify two DNS hosts to be responsible for DNS management of the domain name).

Operation of DNS

When we set up the IP network environment, we usually tell each host the address of the DNS server (we can set it manually on each of the hosts, or we can use DHCP to set it).

Now let's see how DNS works.

When asked about the host name within this domain, the DNS server responds directly
If the queried host name belongs to another domain, it checks the memory to see if there are any relevant information.
If not found, it will be turned to the root server query
Then the root server notifies the domain name's authorized (authoritative) server (which may be more than one)
The local server then queries one of the servers and saves the list of servers in memory for future needs (save the steps for further root queries)
Remote server response Query
Respond the query results to the customer and store the results in a backup in your cache memory.
If you receive the same query before the time of storage is out of date, respond with the data stored in the cache memory.

From this process we can see that none of the DNS hosts that contain all the domain names are distributed in all DNS servers, and the NIC only needs to know the DNS server addresses.

To better understand the operation of DNS let's use the following figure to see the process of querying www.home.netman.com.tw this primary address

In this example, the DNS counterpart for the www.home.netman.com.tw host is managed by the DNS server responsible for home.netman.com.tw this domain name. (In DNS terminology, we call a domain name "zone" zone can be a domain name that you request from a NIC or a "Sub-zone" extended from the domain name). There must be a file on home.netman.com.tw this zone on this DNS server and there must be a record of the WWW in the file (any hosts are represented as "records"). This record can be an IP address or alias to a host name, but in any case the corresponding host name is ultimately to be a corresponding IP address.

DNS also provides the reverse lookup feature, which is to query the host name with IP. Many of the services on the Internet, such as FTP, IRC, www, and so on, require this functionality. In fact, the DNS service itself must use the reverse feature and must also set up a zone for each network. Although some people find that the DNS service can be exploited even without reverse zone, the drawbacks are not easily perceived in this (Chinese) web page http://ns.nctu.edu.tw/Basic/ Whentouse-rev.html above you can see some problems caused by ignoring revers zone.