As the fantasy of superfast internet access draws closer, we can start to see what broadband really looks like

As the fantasy of superfast Internet access draws closer, we can
start to see what broadband really looks like
In the first months of the new millennium, a large blimp will rise silently from a privateairstrip in southern England, drift upward for several miles, then begin beaming high-speedInternet signals to a laboratory below. Big as it is, the blimp will be only the small, initial testmodel for the fleet of 250 vast, solar-powered Internet airships—each a bit wider than afootball field and almost twice as long—that Sky Station International plans to send uparound the globe. Beginning in 2002, if all goes well, Sky Station blimps will float 70,000feet above cities in Europe, Asia, Africa, and the Americas, transmitting Web pages tosubscribers below.
Headquartered in Washington, D.C., Sky Station — the world's leading stratospheric
platform company, according to a company press release posted on its Web site — has been
led since its inception by Alexander P. Haig, the former U.S. secretary of state. The
company's plan, says Y.C. Lee, senior vice president for system technology, is nothing less
than to provide the cheapest, fastest, and most reliable Net access on the planet.
Sky Station will have plenty of competition. A bewildering profusion of cable-televisionfirms, telephone companies, and Internet service providers, many in newly consummatedalliances, are scrambling to deliver high-speed Internet access to the home. Broadband, as thisrapid signal is known, has become one of the Net's most commonly used buzzwords. Andlittle wonder: Over the next few years, the world's biggest telecommunications firms willspend hundreds of billions of dollars seeking the Holy Grail of broadband.
What is broadband? The band refers to bandwidth, engineering jargon for the rate at whichzeros and ones flow through a computer network. The broad refers to the width of the digitalpipeline, which is a metaphor for its capacity. Dial-up modems typically send and receivedata at about 50,000 bits per second; the new broadband connections function at up to 10million bits per second—faster by a factor of 200.
According to many Net aficionados, such blistering speeds not only will transform people's
experience of the Internet but also will create nothing less than a revolution in American life.
The fat pipe, as Internet pundits refer to broadband, will eliminate the typical, text-heavy
Web page of today in favor of continuously moving, multimedia environments — every Web
site its own version of Myst. And it will fulfill one of the ultimate digital fantasies:
downloading movies on demand.
More than that, some enthusiasts predict that swift, cheap data transmission will let designers
embed tiny sensors, smart appliances, and miniature computers in people's walls, chairs,
tables, ceilings, and even clothes. Integrated into a seamless, super accelerated home network,
these aspirin-size processors will track, learn from, and anticipate the needs of everyone in the
home. Ultimately, in the view of Mark Weiser and John Seely Brown, two researchers at
Xerox's Palo Alto Research Center, the impact of such ubiquitous computing will be
revolutionary, and as transformative as the invention of writing or electricity.
Could these visions really come to pass in, say, the next 10 years? Until recently, the questionwas abstract; only big companies and universities had access to a fat pipe. Now, though,broadband is beginning to reach American homes. The first results suggest that it will indeedchange the way people use the Net—but in ways not necessarily envisioned by the experts.
IN SEARCH OF WIDER PIPES
At present, the most widely available means of fast Internet access is the cable modem. Cablemodems connect on one end to users' existing cable-TV feeds and on the other end to specialpieces of hardware called Ethernet cards that are stuck into slots inside computers. Typically,service is priced at about $40 a month, not counting start-up costs. Installation, which iscomplicated, usually involves waiting for technicians; many older cable systems will requireextensive retrofitting. Luckily, Internet usage does not interfere with regular cable-TVservice; one family member can watch the box while another checks e-mail. With cablemodems, Net access is constant. Subscribers are always online.
In theory, every one of the 67 million cable-TV subscribers in the nation should be able to runa cable modem. This prospect makes the cable companies drool—and panic. The reason isthat every cable subscriber in a neighborhood shares the same line. As the number ofneighbors on the Net rises, performance degrades. If two or three people on your street startdownloading videos—boom! You'll know it, says Emily Meehan, a researcher at the YankeeGroup, a technology consulting firm in Boston. To upgrade their systems for broadband,cable firms are already planning to spend $33 billion in the next two years, according to theNational Cable Television Association. If cable modems become too popular, they may haveto spend considerably more—a daunting prospect.
The cable system effectively plugs every computer into a single network with all the othercomputers in the area. When connected, previously isolated PCs suddenly become much morevulnerable to intruders; that's a big problem if, as is predicted, broadband induces morepeople to do their banking and shopping on the Web. Microsoft assumes that anyone who canrun a program can reformat the hard drive, Bruce Schneier, head of the computer-securityfirm Counterpane Systems, has noted. This might have made some sense in the age of isolateddesktop computers; after all, if you could run a program, you were standing in front of themachine. But on the Internet, this is absurd. Cable modems, in other words, could become themeans by which Joe and Joan Six-Pack are introduced to the computer underground.
These difficulties give hope to cable's competition: the telephone companies and Internetservice providers, which are working out the details of a group of related broadband methodscalled, generically, DSL — digital subscriber line. DSLs divide the signals on standard phonelines into a low-frequency channel that carries voice signals and a high-frequency channelreserved for data. Until recently, this required installing a splitter on the user's end, but lastJuly the International Telecommunications Union accepted a new, somewhat slower versionof DSL — confusingly dubbed G.lite — that doesn't need such a device.
Like cable, DSL lets home surfers maintain constant access to the Net. Similarly, telephonecalls over the low-frequency channel are not affected by activity on the high-frequencyInternet channel. Unlike cable, DSL doesn't pool users into a network. Each subscriber gets asingle line, greatly reducing security risks. But DSL is expensive—at least $60 to $80 amonth. (The special modem can cost $200 or more.) And DSL has a limited range: Anyone who lives more than about three miles from the local telephone company's central office isout of luck.
DSL also faces institutional economic barriers, as the Yankee Group's Meehan points out.
Every DSL connection a phone company installs means one less chance to install a secondregular phone line for dial-up access. They're cannibalizing their own revenue stream,Meehan says. You can see why they'd move slow. But the phone companies are being driventoward broadband by the Internet service providers, who are locked out of most cablenetworks.
EarthLink, which will become the nation's second-largest Internet service provider followingits impending merger with MindSpring, is leading the industry's push toward broadband.
According to Jon Irwin, head of the company's broadband division, eliminating the slow,kludgy process of dialing up the Net is essential to getting more people online. Before we hadbroadband in the house, he says, my wife never got on the Net — she just wouldn't do it. Butshe got on when we got fast access, and now she's on it all the time.
Irwin acknowledges that EarthLink faces formidable obstacles in its quest to providebroadband. Cable is in the lead, and DSL is so new that its technical standards are still beingworked out. Nonetheless, he says, EarthLink hopes to have considerably more than 100000broadband subscribers by the end of 2000, many of them on DSL lines in a partnership withthe telecommunications company GTE.
The slow emergence of fat pipes gives latecomers like Sky Station, which won't launch its
fleet of blimps for a few years yet, a chance to catch up. Indeed, the sky will be full of late-
coming broadband providers. If all goes according to plan, Internet aircraft from Angel
Technologies of Saint Louis will be circling nearly 18,000 feet below the Internet blimps
from Sky Station. Beaming broadband to subscribers, Angel's High-Altitude Long Operation
(HALO) aircraft will roam 24 hours a day over metropolitan areas, each acting as the focus of
a Cone of Commerce 50 to 75 miles in diameter. According to a company spokesperson,
Angel will begin deploying HALOs at the end of the year 2000.
Meanwhile, the electronics giant Hughes is planning to devote two satellites to broadband inNorth America alone. An early version of its forthcoming satellite service, DirecPC, has beenavailable since 1996. Subscribers request Web pages over standard phone lines; in RubeGoldberg fashion, the request is diverted to a special, high-speed Hughes network, whichshoots the requested data to a satellite, which in turn broadcasts it to a small dish antenna inthe subscriber's home. Other satellite services are planned by Lockheed Martin, Alcatel, andLoral.
But the most ambitious broadband scheme is from Teledesic, a start-up founded by zillionaire
cell-phone tycoon Craig McCaw and backed by, among others, Bill Gates and Saudi prince
Alwaleed Bin Talal. Beginning in 2004, the company will employ rockets of the type long
used by NASA and the Soviet space program to launch a constellation of 288 satellites,
creating what it calls the first global, broadband Internet-in-the-sky. Total cost: more than
$9 billion.
Because each method of laying a fat pipe faces its own set of obstacles, no one knows whichone will prevail. But it seems implausible to suggest that all will misfire; too much money is being thrown at the wall to suppose that nothing will stick. If so, broadband will be widelyavailable, even common, by 2010.
It'll be ready, all right, says David Nash, director of Intel's Connected.Home initiative, which,among other things, studies consumer broadband use. The only question in my mind is whatpeople will be doing with it.
WHAT PEOPLE WILL BE DOING WITH IT
Next month, workers will finish construction on a two-story house in Atlanta. Architecturally,the structure is utterly undistinguished: a turn-of-the-century & style town house, completewith side porches and a front stoop with Greek Revival & style columns. But to itsdesigners—a team led by Irfan Essa of the Georgia Institute of Technology—the half-million-dollar project is nothing less than a laboratory for exploring the way Americans will livetomorrow.
Essa is not interested in what he calls that Star Trek automatic-house thing — lights that clickoff automatically when people leave a room, doors that slide open as bodies approach, giantmonitors that rise from a recess when hands are clapped together. In his view, most of thesetechno-tricks are already proven failures: There's no automatic security system in the worldwith motion sensors that can deal with pets. His team is after something more complex andsubtle—a house, one might say, that not only recognizes Fluffy but also calls for help whenhe throws up on the carpet.
After the carpenters have finished work, Essa's research team will begin installing anelectronic forest of microphones, sensors, cables, and microprocessors. When operational, thehome network will fuse inputs from multiple sources to track family members. Cameras willrecord faces, sending the data to face-recognition software in the household server; sensorsbeneath carpets will track footsteps, keying them to previously stored patterns; microphoneswill listen near appliances, stairs, and other potential trouble points.
By combining information from these disparate sources, the system is intended to be able todetermine, say, if Uncle George knocked himself out falling down the stairs. In theory, thenetwork would do so by combining data on the old man's identity, whereabouts (foot of thestairs), recent history (unusually rapid descent), and current lack of motion. Using similarmeans, the network could ascertain that the family toddler is leaning out an open window. Orit might match up householders' movements and the contents of the medicine cabinet to seewhether they had taken their medicine that day.
Because such systems would permeate their environment, they are often referred to asubiquitous computing. The term was coined 11 years ago by Weiser, the Xerox researcher,who argued that computers would ultimately disappear into the environment, becoming aninvisible support network. By 2008, he said, imperceptible computers will permeate our lives.
One crucial variable is broadband: The home network needs a very fat pipe to function.
Weiser's vision has been embraced by researchers at MIT, CalTech, Apple, the University ofKarlsruhe in Germany, and other places, although the Atlanta project is the only current effortto build an actual working model of a connected house. The technology is rapidly advancing.
But it also seems clear that ubiquitous computing will not be permeating our lives 10 yearsfrom now.
To begin with, such systems pose formidable programming challenges. We want pencils todigitize what we write, but not if the baby is using them to scribble on the walls. We want themedicine chest to remind us to take our medicine, but not while we're on the phone with theboss. We want computers to call 911 if Uncle George falls down the stairs, but not if he liesat the foot of the stairs and pretends to be sleeping to amuse his young nephews. And so on.
The problems are endless.
Worse, such systems pose enormous inherent risks to privacy. If my house is being monitored,Essa says, I don't want the data to get into anybody else's hands. Yet if we try to lock theinformation inside the house, it means that people lose the ability to, for instance, monitorfrom outside what their elderly parents are doing. Yet that would be one of the best reasonsto have this kind of system.
Given these obstacles, ubiquitous computing seems unlikely to reach the mainstream soon.
But what about the predicted wave of immersive Web sites? In some sense, the question isequivalent to asking whether the virtual exhibit that the Smithsonian Institution is launchingin January is, as the museum hopes, an augur of the future.
THE FUTURE'S SO BRIGHT: The Smithsonian's broadband showcase Sponsored by the Smithsonian's Millennium Project, the exhibit recounts the story of some ofthe museum's treasures, seamlessly mixing music, video, art, morphing images, and 3-Dgraphics. The imagery is astonishing. According to Peter House, webmaster of the virtualexhibition, it is also hard to pull off; few people know how to put together graphicallysophisticated sites. The kind of stuff we have to do, most webmasters don't have anybackground in it, he says. We're taking people out of the community that builds games andCD-ROMs.
As House notes, the shortfall will eventually be made up as programmers learn how toassemble these sites. But they will always be expensive to create, in the way that films areexpensive to create. And to navigate them will always require an investment of the user'stime—something that people may be increasingly reluctant to give in an ever-faster world.
For a lot of our customers, the Net is like the world's biggest and best Yellow Pages, says avice president of one large Internet service provider. They love it, but I'm not sure how manypeople want to immerse themselves in the Yellow Pages.
Researchers at Intel's Connected.Home project largely dismiss the importance of multimedia.
Last spring, the project provided its own example of what broadband might mean, at least in the short run. In a six-week experiment, it provided a DSL connection to 10 families in LakeOswego, Oregon, a suburb of Portland. Along with broadband access, it gave all the familiesa tablet. With their flat screens and simple buttons, the tablets resembled big, pink PalmPilots. Tablet screens depicted a Web browser; the controls permitted the user to manipulate acursor and jump between Web sites. Like a telephone, the tablet was left in the living roomfor people to use as desired. Each family member picked up the tablet, logged on with a touchof a button to an individually configured home page reflecting his or her interests, and thenexplored the Internet. An Intel team of engineers, sociologists, and anthropologists observedthe results.
People's behavior with the Internet really changed dramatically, says Nash, who worked on
the project for Intel. Instead of sitting down and surfing the Net for an hour or so, people
would do what we called drive-by shootings. They would get on the Net for a minute or two
at a time—but do it a hundred times a day. By providing fast access and a simple tool for
using the Web, Nash says, Intel removed all the barriers to full use of the Web. They were
using the Net for frivolous or trivial things. It was so completely painless [to log on] that
people used it to check traffic, the weather, what Johnny's school lunch is, or what the family
needs for dinner. Broadband, he says, can make reaching for the Net just like reaching for the
TV remote control or a cordless phone.
It is in this humble but useful way that Nash thinks broadband will have its greatest impact.
Rapid access will convert the Web into a universal repository for daily knowledge. In thisway, the Net will become a version of what some of the original cybertheorists imagined: auniversal library. Except that instead of using it to consult Plato, people will be using it forschool-lunch menus and synopses of tonight's episode of The Simpsons. That's OK, Nashsays. People need school-lunch menus.

Source: http://www.ai.univ-paris8.fr/~ga/Public/broadband01.pdf