|N E G R O P O N T E||Bandwidth
Even after several generations of being digital, the basics of bandwidth still cause all kinds of confusion: Where does it come from? How much do you need? What does it cost?
It's hard enough to understand without the detours (sorry) caused by likening bandwidth to a highway - the onramps, the offramps, not to mention the roadblocks and tollgates. Rest assured that adding an ISDN or cable modem "fast lane" for your home PC does not solve all the problems.
A more appropriate, if less concrete, likeness might be paranoia, because real and perceived bandwidth are widely separated, and you cannot tell how much of the problem is your own fault. The real frustration comes from the inability to speed up the process by taking any one or, for that matter, any number of measures. The World Wide Wait, as it is too often known, is a chain of many events mostly outside your control, the slowest link of which determines the verve of your connection.
Worse, the slowpoke in the connectivity chain is hard or impossible to identify. Imagine waiting for a bus, not knowing how many people are in line, where you stand in that line, when the bus is coming, or how big it will be when it does. (Woops - just skidded into another roadway metaphor.)
One of the best ways to deal with bandwidth is to understand it on its own terms - which is not easy.
Not as advertised
Bandwidth is the capacity to deliver bits, typically measured by how many you can transfer in one second. Newcomers often don't know the difference between bits and bytes - there are 8 bits in a byte, which happens to be enough to represent a single ASCII character, including standard Latin alphanumeric characters, punctuation, and most accents.
Without going into the brutal details, suffice it to say that you would like any string of 1s and 0s you transmit to be the same as those that are received. This cannot be blindly guaranteed without spending some of that same bandwidth to deliver extra bits for the sake of checking, correcting, or, in the worst case, requesting that bits be resent. When it comes to bandwidth, you're not getting all you think you are - a bit like the coverage of an insurance policy.
But that's OK, if you listen to the press, because a lot more bandwidth is coming, though nobody is sure exactly when. The 16 million-plus miles of optical fiber found in the US alone will soon have the capacity to carry 400 billion bits per second, thanks to recent technology from Lucent (AT&T's former hardware house). The telex, by comparison, operated at 75 bits per second.
All in good time
While 400 billion bits per second sure sounds fast, think of the following: Each CD you own contains roughly 5 billion bits. Imagine a warehouse full of CDs loaded onto an 747 and flown from New York to Washington, DC, in an hour. Guess how many bits per second that is.
Those "Boeing bits," of course, are not accessible in real time. And so? Do you really need all that bandwidth? Do you need it continuously, instantly, versus in a few moments? I often liken bandwidth to a restaurant, where slow service may not matter if you are in good company.
Understanding bandwidth needs is further complicated because bit requirements are all over the map. Words are thin and video is fat: You read (the Latin alphabet) at about 600 bits per second and you watch television at about 3 million bits per second. A picture is not a thousand words, but more like a million.
All bits aren't created equal
After countless hours (and untold billions of bits) of media coverage of the first O. J. Simpson trial, the verdict was rendered with only one bit: guilty or not guilty. I'm sure O. J. values that bit more that any other in his life.
If you have a pacemaker, which sends a few bits each hour to the hospital, there is no doubt that you value those bits more than any of the trillions in Titanic. Not all bits are created equal. But who determines their value, not to mention their priority?
Right now, on the Web, you are more or less subject to the whim of any site. If the site has sold an advertising banner that greets visitors right at the frontdoor, you may have to swim through hundreds of thousands of winking and dancing bits before you get the measly 250 words (an old measure indeed) you were looking for, which alone would have taken about two seconds on your friendly 28.8K modem.
I remember graduating from a 110-bits-per-second modem to 300 bps and feeling the astonishment of speed. Later, 1,200 bps felt like a miracle and 9,600 was lightning. But thereafter, for me, it came to a thud, though I continued to get more bandwidth - there was even a time when I had more Internet bandwidth coming into my home than the entire nation of Switzerland.
That fat pipe now feels empty simply because other forms of congestion get in the way. Communications software has become obese - the comfort of the saddle has separated us from the horse. Intercountry links are slow, often purposely so, because the self-interest of governments or telcos (often the same) are not well served otherwise. Servers are swamped, because bandwidth is also an issue inside a computer, as in the speed at which a processor can talk to memory.
Bandwidth as clean air
The big issue in the near future will be how to charge for bandwidth, if at all. The world of atoms would have us think that large and heavy packages traveling halfway around the world should somehow cost more than a tiny, featherweight object crossing the street. But in the world of bits, that is not necessarily so. I am willing to pay a great deal more, for example, for a few pacemaker bits getting to my local hospital than for receiving CNN in Kathmandu.
At first glance, decoupling bandwidth from value is a nightmare for the telecommunications industry. After all, it costs real money to build the wired and wireless infrastructure needed to wrap the world in unlimited bandwidth. For this reason alone, it will be a while before bandwidth is priced like clean air. Until then, we will all be using bandwidth like scuba tanks. Web designers take note: It pays to be parsimonious.
Next: The Future of Retail
[Back to the Index of WIRED Articles | Back to Nicholas Negroponte's Home Page | Back to the Media Lab Home Page]
[Previous | Next]
[Copyright 1998, WIRED Ventures Ltd. All Rights Reserved. Issue 6.06, June 1998.]