by Tom Snyder on Nov 30, -0001


This month, I continue my series that will attempt to teach you some of the basics of the Internet. This month the subject is How the Internet Works.

The Internet is a loose association of thousands of networks and millions of computers across the world that all work together to share information.

Think of the Internet as an electronic version of a mass transit system in a big city, with a few main subway lines that intersect at certain points. Connecting to the subway lines are commuter rails, bus lines, and ferry boats that spread out and crisscross the metropolitan area.

On the Net, the main lines carry the bulk of the traffic and are collectively known as the Internet backbone. The backbone is formed by the biggest networks in the system, owned by major Internet service providers (ISPs) such as GTE, MCI, Sprint, UUNet and America Online’s ANS.

By connecting to each other, these networks create a superfast pipeline that crisscrosses the United States and extends to Europe, Japan, mainland Asia, and the rest of the world.

When you access a Web site or send an e-mail, that electronic transaction hip-hops through all the networks between your computer and the server that houses the Web site or e-mail server.

To get a graphical picture of this, go to http://www.neotrace.com/ and download an evaluation copy of Neotrace. This great little utility performs a traceroute, which allows you to enter a Web site address and see in table format or on a map all the “hops” your connection makes, beginning with your ISP and listing every node it goes through, ultimately ending up at the target domain name.

While you may think the logical path will be a straight line, that is not always true. For instance, if I do a traceroute to www.chicagotribune.com, my path begins here in Milwaukee, but makes 21 hops, going first to Chicago, but then going to New York, then Palo Alto California, before returning to Chicago and the Web servers that house the Tribune’s Web site.

What that means is, although redundancy can route around trouble spots, there are a lot of things along the way where something could go wrong. If there is only a single route to a origin and destination city, you’ve got a real problem.

For a long time, there was only one Internet pipeline into Milwaukee, and it came from Chicago. If a fiber-optic cable in Waukegan was cut (which happened several times) it would be impossible to for anyone outside of Milwaukee to get to sites hosted here. It would also be impossible for people using a Milwaukee based dial-up to access sites outside of Milwaukee. There was even a period right after the largest local ISP changed its provider, forcing a path to most Milwaukee Web sites to have to go to Chicago and come back to Milwaukee. A cut cable in Deerfield resulted in an inability to access a site hosted across the street!

The mor al of the story is that when you are unable to access a site, the problem may not be the site or the server that hosts it. It may be the path. But it can also be a problem with a Web site’s physical connection to the Web.

In the case of the sites we host here, we have had a 99.9995% uptime over the past two years. However, there have still been occasional rep orts of outages. The good news is that those outages were path problems, or glitches at Time Warner (our upline provider) or Ameritech (the local Telco that’s responsible for the actual wires that our Internet connection uses). The bad news is that those kinds of problems are something neither you or we have control over.