Discovering Ideas

The Internet -- What is it? Where is it?

Let's start with a simple question: where is this page, the one that you're reading right now? In a sense, it's on the surface of your monitor. After all, that's where your eyes are pointed, and without the monitor you couldn't be reading it. But on the other hand, if you turn the monitor off and on again, or replace it with a different one, the page will still be there. What the monitor shows is just a representation, an image of information that is stored somewhere else. So where is the information?

Well, if it's not in the monitor, how about the computer? Certainly, many things you read on your monitor are stored in the computer. Documents you've written or downloaded are stored on your hard drive or on a floppy; any information you've bought in a computer store is probably stored on a CD-ROM (though, while you're working with it, it has to be copied into RAM {C}). But this page doesn't come from any of those places. There is a temporary copy on your hard drive, in your browser's cache, but that's just a backup and will never be used unless you follow a link away from this page and then come back. In fact, if you had a truly permanent copy of this page anywhere on your computer, you'd be violating the copyright agreement that lets you read it!

Legal threats aside, though, if this page is not in your computer, where is it? Let's trace the path back from your screen. The signal to your screen comes from a video card in the computer, which gets its instructions from your CPU (central processing unit). The CPU reads data from the computer's RAM. How did the information get into your RAM? At this point the story changes a little bit depending on your specific system, but chances are it got there in one of two ways: through a modem connected to a phone line, or through an ethernet port. And from there, it's easy: on the other end of the line is Britannica's computer. Right?

Well, no. And here's where we begin to see what's so special about the Internet. In most cases, there is no direct connection between the computer that sends some information and the computer that receives it. Instead, the information is broken up into bits and sent through a series of other computers, across communication links that may be based on fiber optic cables, phone lines, or even satellite transmissions. This is what the Internet is really all about: Relaying messages across a chain of intermediary computers and connections, regardless of their types or physical structures. Any computer that has the appropriate software and hardware to participate in this relaying system can communicate with any other such computer (at least in principle), and can thus be considered a part of the Internet.

From its humble beginnings as a small military network called ARPANET, the Internet has grown to include millions of computers, of all sizes, all over the world. At its heart are the giant routing machines used by the biggest service providers, which do nothing but handle Internet traffic, and without which almost all long-distance Internet connections would be severed. At the outer edges are computers like the one you're using right now, which serve the needs of a single user and could be turned off at any time without affecting the Net at all. And in between are all sorts of networks, corporate and academic, private and public, large and small, which handle not only their own traffic but the traffic of other networks to which they're connected.

If you had a map of all these systems, it would look like some giant rope net or cobweb (which is why we call it the Net or the Web), with points representing computers and lines representing the connections between them.  This structure is referred to as the architecture of the Internet. There are many parallels between this sort of architecture and the sort of architecture that's represented on the blueprint of a building.

Both of these maps tell you how many basic storage spaces (computers or rooms) there are, how big they are, and what path you have to follow to get from any one to any other. In both cases, the architecture defines the physical relations among the parts.

But there's another sort of structure that's equally important to the definition of the Internet. After all, long before the Internet existed, there were networks of computers; before personal computers became common in schools and offices, the standard computing arrangement was a set of simple terminals connected to a central mainframe. But the big difference is that these systems were all isolated. Each type of computer had its own unique way of communicating, and so it was extremely difficult to pass information between them. Part of what made the Internet possible, then, was the development of a set of communications protocols, standards that define a common "language" for communication among different types of computers. Those protocols can be built into any hardware or software, so that even the strangest of devices can now be hooked up to the Net.

Once these protocols were established, it became possible (in principle) for any two computers to share information, so long as there was some chain of physical connections leading from one to the other. To make this possibility a reality, the last thing that needed to be invented was a routing system -- a system that finds the best route for a message to follow. Under the system that was developed, every machine is assigned a unique Internet Protocol (IP) address, a "name" which all other computers on the Internet can use to refer to it. When you send email, or try to connect to another machine using a program like telnet, your computer sends your message as one or more packets, little bundles of information. Each packet has a header, which includes (among other things) the IP addresses of the sending machine and the destination machine. These packets go from your computer to a central network computer, or router, owned by your school or company or ISP (Internet service provider). The router reads the headers (but not the messages!), calculates the best network path to your destination, and sends the packets out to other routers along one or more of its physical connections. This process is repeated until the packets reach the destination computer. The packets that make up a single message may take different paths and arrive at different times, so the headers also carry information about how the packets should be reassembled into the original message.

While this system may seem rather baroque, it allows for trememdous flexibility: If a computer or network crashes, it doesn't disrupt the whole Internet, because the routers connected to it just block it out and find alternate routes for the messages that they would normally send through it. Likewise, if you start up a new network, you don't have to tell every computer on the Internet about it; you just tell the router that you're connected to, and it tells a few of the biggest routers (often called backbones), and that's enough for any message to find you.

This flexible, abstract structure is called the virtual architecture of the Internet. The virtual architecture determines which computers can possibly communicate with each other, no matter how many others they have to go through to get there. The reason it's called "virtual" is that it has to do with what seems (to the user) to be connected, rather than with what really, physically, is. When you're engaged in real-time Internet chat with a friend in Venezuela, it seems as if the computers on your desktops are directly connected; likewise, if another computer is physically connected to your network but it's behind a firewall or not running any Internet protocols, it may seem as if you have no connection to it. Together, the virtual architecture and the physical architecture also determine the speed and reliability of communications; if your friend next door has a slow modem and subscribes to a different Internet service than you, it may take longer to get a message from them than from an e-pal in France who uses the same service as you and has access to a T1 line (a high-speed digital connection).

Now we can finally come back to our inital question, and give it a more sophisticated answer. Where is this page? Well, it "lives" on Britannica's computer, since that's the only place that has a permanent copy of it. When you clicked the hyperlink that took you here, a copy of the page was sent from Britannica's computer through a bunch of routers, and eventually into your computer's RAM. You're reading a display of that copy, and there's also a temporary copy on your hard drive, in your browser's cache.

Of course, there's still something funny about all this "copy" talk. We say that "a" copy was sent from Britannica's computer to yours, but couldn't we say that it was copied many times? After all, when a router is receiving a page it has to store it in RAM, and when it sends the page along it doesn't somehow "squirt it out" of RAM -- it sends a copy of what it has, and then deletes it from its memory. So in principle, we could say that the page got copied once per router, meaning that you have something like a tenth-generation copy of the original page. But then, what about your modem or ethernet card? In a sense, pieces of the "copy" have to be assembled there before they're passed into your computer's RAM. So is it one more copy per modem or Ethernet card? How about the communication lines? During transmission, is a "copy" of the page somehow stored in the line?

What all this is suggesting is that information is a very odd sort of thing, once you start to look at it closely. We'll deal with such issues at much greater length in later sections, but for now, just let some of these questions percolate in the back of your head (don't worry, it doesn't violate the copyright agreeement for you to keep a copy there). In the meantime, now that you understand the structure of the net, you can move on to the next section, which deals with its contents.

Discovering Ideas
Palomar College
jtagg@palomar.edu
This page was last edited: 08/17/04