Previous StoryNext StoryTable of Contents



Post-Text











































If someone were to chart a map of the Internet, it would look a lot like a page out of a road atlas. Just as a network of roads and highways link cities together, a dense web of wires and cables connects the network information centers throughout the world. When you send an electronic request -- a simple e-mail message, a file transfer, or a URL request -- it travels along the Internet as packets of information. It's possible, albeit tricky, to track the path of those packets traveling from your computer to the final destination.

Traceroute is a utility long used by users of UNIX operating systems to track the progress of information packets on the digital highway. It is intended for use in network testing and management by network specialists. But given some understanding of how it works, even average Web surfers can glean information from Traceroute. Below, we show you how.

Hopping Along the Internet

Every computer connected to the Internet has a host name (represented in its e-mail address or a URL). Each host name has a corresponding numerical identifier, typically four numbers separated by dots (e.g. 233.45.128.211). Information packets travel or "hop" from one information transfer center to the next through the complex circuits of the Internet, on the way to the destination computer.

The first hop is usually from a user's host computer to the user's Internet Service Provider (ISP). The ISP then launches the packet on a sequence of hops to information transfer centers, also known as public exchange points, along the Internet. At each public exchange point, the packets of information pass through a router. The router reads the destination IP number and determines the most efficient path on which to send the information packets.

Traceroute software follows the packets of information on their journey through the cables, routers, and transfer stations that make up the Internet, and produces an onscreen record of each hop.

How does it work?

When a user does a Traceroute, a series of special packets of information is sent out from the user's computer to the destination computer. The first packet travels just one hop, the second two hops, and so on. As each of these packets in turn reaches a new node on the route to the destination computer, successive routers send back responses. These responses in effect "trace the route" of information sent between the two locations. For example, a Traceroute from the PreText server to MIT in Cambridge, Massachusetts might look like this:

Traceroute to mit.edu (18.72.0.100), 30 hops max, 40 byte packets.

1 199.238.225.129 (199.238.225.129) 1.101 ms. . .
2 cafe-gw.speakeasy.org (199.238.226.100) 2.718 ms. . .
3 wes-border2.nwnet.net (204.200.245.54) 3.572 ms. . .
4 wes-core2.nwnet.net (204.202.44.34) 6.814 ms. . .
5 204.202.47.20 (204.202.47.20) 4.076 ms. . .
6 sea1.pdx1.verio.net (205.238.56.29) 14.497 ms. . .
7 pdx1.pdx0.verio.net (205.238.56.53) 9.016 ms. . .
8 pdx0-3.paix-3.verio.net (205.238.56.50) 49.077 ms. . .
9 paix.paix1.verio.net (205.238.56.90) 55.484 ms. . .
10 paix.bbn.verio.net (205.238.52.138) 49.98 ms. . .
11 su-bfr.bbnplanet.net (4.0.1.49) 40.131 ms. . .
12 paloalto-br1.bbnplanet.net (4.0.2.194) 30.242 ms. . .
13 chicago1-br2.bbnplanet.net (4.0.3.165) 81.941 ms. . .
14 boston1-br1.bbnplanet.net (4.0.1.126) 85.79 ms. . .
15 boston1-br2.bbnplanet.net (4.0.2.250) 119.119 ms. . .
16 cambridge2-br1.bbnplanet.net (4.0.1.186) 116.385 ms. . .
17 ihtfp.mit.edu (192.233.33.3) 133.413 ms. . .
18 E40-RTR-FDDI.MIT.EDU (18.168.0.11) 99.979 ms. . .
19 MIT.MIT.EDU (18.72.0.100) 97.472 ms. . .

The above Traceroute report shows the route taken by data sent from our server to the central server at MIT. The packets sent from the PreText server successfully hop--or pass through all of the routers--at 19 public exchange points. In all of the steps except the first, the location of the hops can be determined by their host names. Step #1 on the list represents the server from which the Traceroute was sent. (This is not the main PreText server; it's one we had to borrow for this demonstration). Given the "sea" prefix listed in line six, it's safe to assume that the packets remained in Seattle through the sixth hop.

"pdx", seen in hops seven and eight, is a well-known abbreviation for Portland, Oregon, so it's likely the request passed through two routers there. Moving down to number 12 we see a host name with the prefix "paloalto". In fact, the Digital Internet Exchange, a highly trafficked national information transfer center, is located in Palo Alto, California, so we can be fairly certain our packet passed through there. From Palo Alto the packet was sent to Chicago, then Boston, then Cambridge before passing through two intermediary routers at MIT to reach its final destination: the central server at MIT.

Delivery Problems

Just as an airplane can only fly a certain distance before running out of fuel, a packet of information can only pass through a certain number of hops before reaching its final port. If it's forced to go through too many different routers or is delayed at certain public exchange centers the packet will be dropped or will "die" -- meaning the packet delivery has been terminated.

If there's a delay, or even a complete packet loss, a network manager can determine where the glitch occurred by the way the message and IP number -- or lack of message and IP number -- appears onscreen. If a packet doesn't go through at a particular hop, Traceroute will show a series of asterisks after that hop's line number before dropping off. Electronic traffic congestion, technical error, a bug-ridden router, or a broken connection may cause similar Traceroute results.

Considering the millions of packets of information that traverse the Internet, it's amazing that more aren't lost. For most glitches the Traceroute readout can help identify the problem. A network manager can see if a reported problem is inside or outside the local network, and decide whether to simply send the request again or contact someone at one of the public exchange centers for help.

Tracing Your Own Route to PreText

We've set up a Traceroute program that lets you see the hops between your server and the server that hosts PreText magazine. Step #1 on the Traceroute will begin with the PreText server and should end with the host name and/ or binary number for your own computer. Be warned, it may turn out more complicated than the example above. Some network administrators bar Traceroute requests because they think that revealing binary IP numbers associated with their intermediary routers poses a security risk. If you are working off such a network, or for a company with sensitive security concerns, the Traceroute will go through, but the routers within your organization will be represented onscreen by asterisks. And if one of the legion of possible problems mentioned above arises en route, you may find Traceroute spilling out some impenetrable gobbledy-gook.

But if you are lucky, and the journey from your computer to ours is straightforward, you'll be able to read off the stops along the way. Just click here.


___________________________________________


by Susan Dumett


___________________________________________


Send your e-mail feedback to Susan Dumett.
Selected comments may be published on our "E-mail to the Editor" page.




Previous StoryNext StoryTable of ContentsBack to Top
About PreText


    � 1998, PreText, Inc.

    PreText Magazine is produced by PreText, Inc., a digital media studio specializing in content, design, and technology for the Web. For more information visit the PreText Web site.