|[ Team LiB ]|
Domain Name System
Although DHCP does a good job of hiding your own computer's IP address from you, it does nothing to make IP addresses manageable on the Web. So that you don't have to type into your browser the IP address of each computer you want to visit on the Web (you can if you want, by the way), the Web's creators developed the Domain Name System (DNS).
From your perspective, DNS works by assigning a structured name to every Web site, the familiar whatever-dot-com you use every day. In the language of the Internet, that dot-com is called the domain name of the Web site.
Most domain names take the form of a word, a period, and another few letters. Those letters after the period comprise the top-level domain, the primary organizing structure of the Web. You can't use just anything as a top-level domain. The organization charged with administering Web names, the Internet Corporation for Assigned Names and Numbers (ICANN), maintains tight control of top-level domains.
There are two kinds of top-level domains: organizational and national. Originally, only six organizational top-level domains were allowed, but on November 16, 2000, ICANN added seven more. Table 14.3 lists the currently recognized top-level domains.
In addition, each nation in the world is also given its own top-level domain, a two-character country code. These are listed in Table 14.4.
To the left of the period is the name of the actual domain assigned the Web site through a registry. If there is more than one period in the name, the leftmost portion is a subdomain of the next domain name to the right. The DNS system allows for multiple subdomains. Each subdomain (or domain, if there are no subdomains) specifies an actual server on the network.
To the right of the top-level domain is the directory path to a particular file on the designated server containing a Web page or other data. In Internet lingo, this composite construction of the domain name and directory path is called a Uniform Resource Locator, because the name itself holds all the information computers on the Web need in order to find a particular page or file.
The best view of the Internet comes with following a packet sent from your computer. When you log in to a Web site, you actually send a command to a distant server telling it to download a page of data to your computer. Your Web browser packages that command into a packet labeled with the address of the server storing the page that you want. Your computer sends the packet to your modem (or terminal adapter), which transmits it across your telephone or other connection to your Internet Service Provider (ISP).
How DNS Works
The instant you press your mouse button with the cursor pointing at a particularly juicy image on your monitor, your computer drops everything to obey your command. The mouse sends your microprocessor an interrupt to make it pay immediate attention. The mouse driver checks to see whether a signal has come in saying you've pressed the mouse button. The driver passes this vital information to your operating system, which takes a peek into its private memory to see what location value it has stored for the location of your mouse's cursor on the screen. The operating system then checks to see whether the mouse location corresponds to a hotspot on the Web page you're viewing, indicating to your browser that there is a hyperlink instruction to send you to a new location on the Web. When there is, the fun on the Web begins.
Your browser has no idea where to find the page you want on the Web. All it has is a name—it's sort of like finding an isolated name on a message pad when you awake from a drunken stupor. You recognize it as a name, but you don't know why you wrote down the name or how to get in touch with the person to find out.
You might start with the white pages of your telephone book, but your computer can't even open a book, let alone look something up. The only thing it can do is send out an electrical signal down the connection with the Internet. It doesn't send out just any signal. It sends out the name of the Web site you're looking for in a special data packet. Rather than the white pages, the name goes to a resolver.
A resolver is not a special machine. Rather, the term resolver defines a particular function of a special kind of server on the network, called a name server. The resolver does what its name says—it tries to resolve the address of a site on the Web. It looks at the name your computer has sent to it in a standard form known as a Uniform Resource Locator (URL).
Like a mailman sorting mail and looking at the bottom of the address first (for the ZIP Code and state), the resolver examines the last part of the URL first—the part of the name to the right of the rightmost period—the familiar .com, .org, or .edu. This portion of the name is the top-level domain, which tells the resolver how to find the location of the Web site. In Internet terms, the top-level domain is about as general as describing a creature as being in the animal kingdom.
The top-level domain doesn't tell the resolver where to look for the Web site. Rather, it tells the resolver where to look for a list of site names in the top-level domain. Resolvers aren't stupid. If they've looked up a top-level domain before, they probably already know where to look. If not, however, they call on one of the 13 root name servers to tell them which servers store the information about each top-level domain.
The root name server, or, more likely, the resolver, passes your request for the Web site to one of the name servers assigned to the top-level domain of the URL you're looking for. Hundreds of thousands of servers may track this information. It's kept in multiple copies for speed and reliability. Speed because one server is not burdened with finding every requested URL in its domain, and reliability because if one server becomes unavailable, there are hundreds of others that can take its place.
This server matches your requested URL with the domain name server (also abbreviated DNS) that handles the Web site. The server sends the requested URL to the IP address—a block of four bytes of binary code—of the DNS. The DNS knows all the names of the Web sites it serves. It passes the IP address of the Web site you want back to your computer so it can use this address to find the page you want.
When your computer signals to the IP address, it sends a request for the page listed in the hotspot you clicked. The server at the Web site diligently finds the page and passes it back to your computer. Your operating systems passes it to your browser, which formats the page for the screen and passes it back to your operating system, which sends it, in turn, to your display driver and then your monitor screen.
All these requests travel from server to server with light speed, so everything happens fast. Your computer should know the IP address of the page you want and start loading the page in a fraction of a second. Meanwhile, you've probably become impatient and clicked on something else, starting the whole process over again.
Root Name Servers
The part of the Web in charge of identifying each site and getting its address to you is called the Root Name Server System. The master plan that makes it work is the DNS protocol, which describes the packets that need to be exchanged and provides the roadmap for them to follow. The root name servers hold the key to locating the indexes containing the IP addresses you need. The root zone file is the index itself.
The root name servers are arguably the most important computers on the World Wide Web. Only they store the official records of the locations of the registries for each top-level domain, the rightmost part of each Web address. Because this information is so vital to the operation of the Web, it's stored not in triplicate but in 13 duplicate copies in separate computers spread across the world.
Well, not quite. The 13 root name servers actually represent only six distinct geographic locations. Six are clustered around Washington, D.C., two are co-located (and co-operated) in Marina Del Ray, California, two are in the Silicon Valley area, one is in Japan, one is in the U.K., and one is in Sweden. Table 14.5 lists the 13 root name servers.
There's another piece to the IP naming system: the names your computers wear when you access them over your own network using the networking capabilities of Windows. You assign these names using the Windows Internet Name System (WINS). This system is responsible for converting the names you assign to IP addresses for routing messages through networks using the Internet Protocol. In effect, WINS works like DNS at the local level.
In the WINS system, you assign your computer a name when you set up networking on that machine. The computer then sends its name to the server, and the server stores the names you assign in a database, which the server references to resolve requests for IP addresses.
|[ Team LiB ]|