[Back to Lecture Notes page]
The Domain Name System (DNS)
Subtopic outline:
The Domain Name System (DNS)
- Users and applications usually operate on addresses in the symbols and strings.
h.hiew@murdoch.edu.au
www.yahoo.com
- Network software, on the other hand operate on numbers.
- There needs to be a way of converting from one format to another.
The Domain Name System (DNS)
- In early network machines
- Each machine has a name and a network number
- The conversion is store in a file on a central machine, and all host download this file every day.
- This is not feasible today because
- the file would be too big since number of machines to be connected have reached millions, and
- The names chosen probably conflict.
- The DNS was created to solve these problems.
How DNS Works
- In a host machine, we have applications which has names to be resolved to network IP numbers.
- In the host machine is also another piece of software called the resolver, which the applications will pass the names to.
- The resolver software sends UDP packets to query a separate local machine called the DNS name server, which contains a listing of name to IP number conversions.
- The name server sends the replies back to the resolver with the appropriate IP numbers.
- The resolver passes the IP numbers back to the applications.
- But what does the DNS name server store? Surely not all possible name to number conversions! That
is what we will consider for the rest of this lecture.
The DNS Name Space
- We divide the Internet into different domains, in the form of a tree (called the DNS tree):
Figure 7-25 p623
The DNS Tree
- In the DNS tree, a domain is a specific node plus all the nodes under it.
- We refer to domains in the "dot" notation, eg. murdoch.edu.au.
- Node names have a maximum 63 characters, and maximum domain names is 255 characters.
- Each domain is under the control of a particular organization – eg. all the top-level country domains are controlled by the countries themselves.
- To create a new domain, permission must be obtained from the controllers of the domain which the new one is to be included.
- The DNS do not need to follow the boundaries of the physical network.
- Eg. we can have two sets of hosts on the same LAN split into two domains.
Resource Records
- Every domain has a set of resource records (RRs) associated with it.
- The most common RR is the the IP address.
- The job of the DNS is to map domain names to RRs.
- When a resolver queries a DNS name server, they get back RRs.
- Domain name
- Time to live
- Class
- Type
- Value
Types of Resource Records
- What kind of information can you get from a name server?
Figure 7-26 p625.
The DNS Database
- Each name server keeps a database of the resource records.
- When the name server receives a query from a resolver, the name server looks up the appropriate entry (or entries) in the database, and returns the RRs.
- An example of RRs stored by a name server:
Figure 7-27 p627
Authoritative Record
- When a resolver receives an RR directly from the name server, the RR is called an authoritative record.
- In contrast, if it is a record retrieved previously, stored up, and now reused again, it is a cached record.
The Name Servers
- To determine which name servers handle which nodes in the DNS tree, we divide the domains into non-overlapping zones.
- The zones do not need to follow the hierarchy of the domains, but all nodes in the zones must be directly connected.
Figure 7-28 p628.
- Each zone would contain one primary name server.
- Each primary name server knows about
- the primary name servers for all its children zones, and
- the root name server of the DNS tree (and some of the more popular top-level domains, like .edu and .com)
Querying Name Servers: How it works
See figure 7-29 in the textbook and read the description of the example on p629.
[Back to Lecture Notes page]