Like Dynamic Host Configuration Protocol (DHCP), DNS is a client/server protocol. This means there is a client component and a server component necessary to successfully implement DNS. Because of the client/server model, any computer seeking DNS information is referred to as a DNS client, and the computer that provides the information to the client is referred to as a DNS server. It's the job of a DNS server to store a database containing DNS information, to respond to DNS queries from clients, and to replicate DNS information to other DNS servers as necessary.

DNS provides for several types of queries, including forward lookup queries and reverse lookup queries. Forward lookup queries allow a client to resolve a host name to an Internet Protocol (IP) address. A DNS client makes a forward lookup using a name query message that asks the host address record for a specific host. The response to this query is sent as a name query response message. If there's a host address record for the specified host, the name server returns this. If the host name is an alias, the name server returns the record for the alias (CNAME) as well as the host address record to which the alias points.

Reverse lookup queries allow a client to resolve an IP address to a host name, as Figure 26-1 shows. Reverse lookups are primarily used by computers to find out who is contacting them so that they can communicate directly using an IP address rather than a host name. This can speed up communications in some cases because name queries aren't necessary. A DNS client makes a reverse lookup using a reverse name query message. The response to the query is set as a reverse name query response message. This message contains the reverse address record (PTR) for the specified host.

Figure 26-1. A reverse lookup query.

DNS also provides a way to cache DNS information to reduce the number of queries that are required. So, instead of having to send a query to a name server each time the host wants to resolve a particular name, the DNS client checks its local cache for the information first. DNS information in the cache is held for a set amount of time, referred to as the Time to Live (TTL) value of a record. When a record exists in cache and its TTL has not expired, it is used to answer subsequent queries. Not only does this reduce traffic on the network, it also speeds up the name resolution process. A record's TTL is set in the query response from a name server.