May 22, 1997
Sky-High Dreams for the Internet
By PETER WAYNER
t first glance, the plan by Craig McCaw and Bill Gates to circle the Earth with 288 satellites supplying high-speed Internet access to the world sounds like a broad grab for power that only a pair of multi-billionaire megalomaniacs could dream up. Closer scrutiny suggests that the plan may actually be feasible -- at least from a business standpoint --even if technical challenges loom large.
Until now, satellites have been produced one or two at a time, with price tags ranging upward from $100 million. Only big governments, black-budget military agencies or huge multinational corporations could afford them.
In This Article
Competitors The Technology Worldwide Businesses Determining Competition Conclusions Related Articles
Wired Truckers: Cowboys of Cyberspace (November 21, 1996)Satellite Fantasies I: 'scuse Me While I Kiss the Sky
(June 12, 1996)Satellite Dreams II: A Rude Awakening to Reality
(June 19, 1996)
In this light, manufacturing and launching 288 satellites seems less like a business plan and more like the dream of a personal electronic Taj Mahal.
In fact, the project being built by the duo's Seattle-based company, Teledesic, is just a follow-up exercise in revising and extending a long-running trend in telecommunications. While the charisma and personal presence of both McCaw and Gates draws plenty of media attention to Teledesic, at least seven other similar endeavors are already under way.
Taken together, the eight plans would bring assembly-line efficiencies of scale to the business of making and launching satellites. In all, least $25 billion may be spent to launch more than 550 satellites providing different levels of wireless Internet and voice access to the world. Teledesic's plan, which is only moderately more aggressive than the others, should cost about $9 billion if the company's estimates are correct.
Deep space is not the only arena for these developments. Numerous terrestrial wireless networks are in development, from extensions of the cellular phone system to entirely new metropolitan-area networks offering bandwidth on demand. One project, known as SkyStation, proposes to use balloons instead of satellites to relay the data. The company plans to build dirigibles that stay aloft for months parked over major metropolitan neighborhoods where demand for bandwidth is high.
The proliferation of businesses moving into this arena is a good indication that the technology is mature enough to support the investment. But, the road for each of these companies promises to be very competitive, and it is far from clear that any one, in particular, has the winning solution.
What's more, many of the systems are still on the drawing board, and there is bound to be a serious shakeout before all are developed successfully. Many are also competing for different niches, and it is entirely possible that several will survive in different roles.
There are also many technological hurdles to leap. Although the basic radio technology is largely proven and even in use today on the ground, many of the practical details of implementation may add unforseen complications. The Teledesic system, for instance, uses very fast frequencies that are often blocked by trees. Farmers in Kansas or strip mall shopkeepers may have no trouble adding an antenna, but many buildings on either tree-lined streets or in dense city neighborhoods may not be able to maintain a clear view of the satellite.
Many of these practical issues will not become apparent until the systems enter their final testing phase. These hurdles guarantee that the marketplace for Internet access is going to change radically over the next four or five years that the satellite networks are deployed. While companies like Teledesic are high-profile, serious efforts, there is no guarantee that they'll be able to beat out any of the other schemes.
eledesic's biggest and most active competitor is Motorola, which launched the first 5 of its 66-satellite Iridium network on May 5. The company is promising to be in service by 1998, but its success at meeting this target will largely depend on launching the satellites.
The Iridium network promises to offer cellular-like phone service worldwide, albeit at what looks like a significantly higher price than standard cellular service (estimates are about $3.00 per minute). Data communications will also be available but at a relatively poky rate of 4.8kps. For this reason, Russell Daggatt, the president of Teledesic, doesn't consider Motorola's Iridium network a direct competitor for Teledesic's high-speed connections. "In some ways," Daggatt says, "they're a technology demonstration for us."
A French company, Alcatel, is also proposing its own network of 64 satellites that would compete directly with Teledesic for high-bandwidth Internet connections. Although Alcatel's proposal for the system, known as SkyBridge, didn't reach the Federal Communications Commission until late February, the company is moving rapidly and has the advantage of being an established company with experience in space technologies.
"Alcatel's built satellites before," noted Phillip Spector, a partner from the law firm Paul Weiss, Rifkind, Wharton & Garrison, which prepared the company's FCC application for spectrum use. This may be why Teledesic recently formed a strong partnership with Boeing and granted the Seattle-based aerospace company the right to launch Teledesic's satellites.
The Alcatel system also plans to offer direct, high-bandwidth connections, but its system will be significantly simpler and thus less expensive, with an estimated price tag of $3.5 billion. Alcatel satellites will only reflect signals from a local ground station known as a gateway and won't have the ability to route data among the satellites themselves. The data that move between the gateways will have to run through another network, most likely the terrestrial fiber network.
The Alcatel system depends upon a fairly unique plan to share the spectrum in the KU band, the area or radio frequencies currently being used by direct-broadcast satellite television systems. Its satellites would shut down whenever they might interfere with these already operating television systems.
Alcatel holds a major regulatory advantage. Teledesic spent valuable political capital getting a block of frequencies allocated to it worldwide, and it was forced to turn to the very high-speed KA band, which was the only region of the spectrum open enough to accommodate anticipated demand. Alcatel predicts that using the lower frequency KU band will make it easier to borrow from the current technology developed for television broadcast.
And there are more plans, schemes and trials in various stages:
- Another major player is likely to be the cell phone giant Qualcomm, which chose Loral as a partner in its Globalcom system, another low-bandwidth system that will be a more direct competitor with the Iridium network.
- Inmarsat, a company that already offers global phone service with a suitcase-sized device, announced it has already raised more than $1.4 billion for its ICO project.
- TRW and Teleglobe Canada are planning a global cellular system called Odyssey.
Most of these plans are much closer to a business plan than a working system, but their proliferation means that the Teledesic project is less a visionary endeavor than just one take on the next step in telecommunications.
Metricom, for example, is aready offering very good low-speed Internet service in several major cities through a network of low-power antennas mounted on lamp posts and other relatively pedestrian spots. Their latest modem is the size of a pack of cigarettes, and while the bandwidth is only that of standard 33.6 Kbps modems, many users are enthusiastic about the service.
Proprietary satellite systems are not new either. Qualcomm introduced very low speed satellite service to truckers in the late 1980's, and its OmniTRACS system is used by almost all major trucking companies to communicate with drivers and monitor performance. The company now offers route planning software and other enhancements that make the communication link a fundamental part of a business that was once very lonely and independent.
The rush to build vast systems of satellites in low orbits -- that is, altitudes of 700 to 1,000 miles -- is just the latest extension of this trend. Despite their expense, satellites might make ideal platforms when compared with other telecommunications technologies.
Fiber-optic cables, for example, must be laid underground by a skilled crew and, more importantly, become a very expensive infrastructure in sparsely populated suburban and rural areas.
A visualization of the Teledesic Network. Cellular transmissions require relatively large antennas that are often opposed by people who either disapprove of their looks or worry about the effects of electromagnetic radiation.
Satellites, in contrast, fly well above all this terrestrial fray.
Keeping the satellites in a low orbit is a crucial part of the scheme. A radio signal travels at the speed of light, but that still takes about a half a second to cover the distance to a geosynchronous satellite -- the kind that orbit the equator at the same velocity as the Earth's rotation and thus stay put in one place in the sky. This half-second adds a large amount of latency to a request for data and makes the design of systems significantly more complicated.
Closer satellites also can focus their signal on a smaller part of the Earth and thus use less power. Both of these effects make it easier to reuse the same spectrum in other parts of the globe.
But the low orbits add other complications. Geosynchronous satellites are named that because they remain directly above one spot on Earth. This significantly reduces the complexity of ground antennas, because they can remain locked on a satellite for long periods of time. This allows direct broadcast satellite television systems to use relatively inexpensive antennas that have a fixed attachment.
Geostationary systems also require only one satellite for a particular region. Low orbit satellites, on the other hand, travel so fast that each of the proposed systems calls for launching large constellations to ensure that one is always overhead. The Teledesic satellites, for instance, will circle the globe every 100 minutes, and Teledesic estimates that an antenna on the ground will have to switch satellites every four minutes.
The rapidly moving satellites offer both electronic and mechanical challenges to the designers. Cellular phone systems already hand off control from one antenna to another as cars move between cells, and this experience should make it easier for Teledesic to design the electronics to hand off between satellites.
But the high frequencies and long distances require more focused antennas. The Teledesic and Alcatel antennas will need to track the satellites as they make their trip across the sky. But since antennas that move physically would be crippled by mechanical failure brought on by dirt and corrosion, all the companies are examining the use of active antennas that use electronics to track the signal without any mechanically moving parts.
What is more ominous, though, is that rapidly moving satellites could pose a big problem to customers who lack a clear view of the sky. The Teledesic system, for instance, promises to have at least one satellite at least 40 degrees from the horizon at all times and will often have more than one accessible. The antennae's view of the satellite must remain clear, however. Any trees in the way will block the signal, resulting in intermittent service.
The social ramifications of this sudden parity is almost impossible to predict because the last technology to offer such a leveling effect was the airplane. It brought remote areas of the globe closer together, but it still required a heavy amount of infrastructure to support the planes.
Today in CyberTimes
ARTICLES AND COLUMNS
Sky-High Dreams for the Internet
By Peter WaynerPanel Sees Holes in U.S. Internet Security Plan
By Steve LohrRussian Women Take to the Web to Battle Stereotype
By Lisa NapoliTamagotchi: Love It, Feed It, Mourn It
By Carol Lawsonarts@large: Listening to the Voices of the Internet
By Matthew Mirapaul
The Iridium network, with its handheld, battery powered handset, is bound to become popular with remote explorers because it guarantees coverage everywhere. Neither Alcatel nor Teledesic will be able to offer as much portability because of the power requirements and sophistication of their antennas, but they'll still be able to offer high-speed connections with a dish about the size of a pizza box.
The requirement for worldwide service represents both a cost and an opportunity. More than 70 percent of the globe is covered with water, and satellites over this region will have nothing to do until they arrive back over a major city. But during this time, Teledesic can also provide service to ships at sea and small islands that may never have had any high speed option like fiber.
Pricing services in all of these locations is certain to be a major undertaking. Teledesic, for instance, is planning partnerships with local phone companies, which will resell the service in each country. This will help Teledesic negotiate the regulatory maze and also isolate them from any complaints that may arrive when people discover that customers in another country are paying substantially less.
The final prices for the service are anyone's guess. It is in each system's best interest to charge what the market will bear, but what this will be may not be obvious. Heavily developed nations like the United States or Japan are much wealthier, but they also have alternative -- and competing -- technologies in fiber networks. Underdeveloped countries, on the other hand, don't have much to spend, but they often have no other option if they want high-speed service.
The SkyStation balloon-based system may be an ideal compromise. They only need to deploy dirigibles in whatever markets they choose to serve and are not bound by technological requirements to provide service anywhere else. Their balloons will not spend 70 percent of their time over water, and this lowers the cost of building the network. It also makes it easier to add capacity. If one city requires more coverage, another balloon could be sent aloft. Systems like Teledesic or Alcatel, however, would be forced to add multiple satellites. To double the number of satellites over New York City or London would probably require doubling the entire number aloft.
ne of the most complicated questions facing any analyst attempting to predict the success of these systems is determining who's competing with whom. On the face of it, Iridium and Teledesic seem like competitors. But, in reality, they may become close friends and or even partners because both may end up serving distinct markets. Iridium users may pay a premium for highly portable but low-bandwidth service while Teledesic will aim at offering high-bandwidth service to fixed sites.
Nor is it obvious that the big trunks of fiber optics spanning continents and oceans will offer much competition for Teledesic or Alcatel. They both concede that fiber-optic networks offer data transmission at prices that satellite systems will be hard pressed to beat. But paradoxically, the biggest competitors for these global satellite systems will be the local telephone companies and their local access network. Both Alcatel and Teledesic acknowledge that they're basically offering a "last mile" solution that provide high-speed access without laying fiber-optic lines.
Many businesses in even the most developed regions of New York City or the industrial suburbs are discovering that getting high-speed access is not as easy as getting the phone company to drag fiber to their building. The cost depends upon the infrastructure already in place in the neighborhood. This is why the easiest place to get high-speed access is often in the major business centers like Manhattan where the phone company has already installed fiber-optic cable to many buildings.
None of the satellite systems has any of these problems. It is as easy to place an antenna on the top of a building in Manhattan as it is to put it on a building in Binghamton, Yonkers, Yosemite Park or the small island of Lanai in Hawaii. In fact, the requirements for a clear view of a satellite may make it difficult to place an antenna on a short building in a neighborhood filled with tall ones.
The competition between users will also make this system a less viable alternative in the cities. Teledesic estimates that each satellite will support about 125,000 basic channels with 16 kilobits of data per second traveling over each channel. Basic users will be able to grab multiple channels for higher bandwidth.
The users in a cell will need to share this bandwidth with all other units covered by the same satellite. But since the number of satellites overhead will change continually, it is difficult to estimate how much sharing will necessary. All anyone can say is that it will be substantial. The entire San Francisco-Oakland area will be covered by one "super cell," roughly 100 miles square. Each satellite will be able to reach and service about 64 supercells at a time, but it will share this load with other satellites that are also passing overhead.
What this all means is that a region like Silicon Valley, with a dense concentration of technically demanding customers, could overwhelm the network, and customers could become frustrated with sharing bandwidth.
A far greater challenge will be getting the pricing correct. No one is certain what demand there will be for high-bandwidth communications, but all the companies planning satellite transmission assume snowballing growth in demand.
Profitability may simply be the difference between making the right technical decisions. Alcatel's system will be significantly cheaper because the company does not plan to add any intelligent switch capability to its satellites. This lower cost could mean the difference between a loss and a profit. Or the absence of this capability could force it to pay too much for ground-based fiber communication to knit together its base stations.
There are thousands of decisions like this to be made in designing the systems, and all the systems' designers are hoping they're making the right call.
There may even be radical combinations that develop over time. A company like Teledesic might find itself a partner with a terrestrial network like Metricom. A neighborhood might be served by one high-speed down-link that then distributes its data over micro-cellular repeaters sitting on lamp posts.
A similar solution was recently used in Mongolia by a National Science Foundation project. The system used a satellite down-link to connect the country with the Internet. Local traffic, however, used small radio links because the wired infrastructure was too old and noisy to support fast signals.
Of course, the forces of time and development will also affect the economics, and the satellites may only be financially viable for a limited amount of time. Peter Huber, a Forbes magazine columnist and the author of several influential reports on the telecommunications industry, points out that the battle between wired and wireless communication has been constantly changing.
"One minute everyone was putting up microwave towers to carry long distance service," Huber said. "And several years later they were dynamiting them as fiber took over."
Related Sites
Following are links to the external Web sites mentioned in this article. These sites are not part of The New York Times on the Web, and The Times has no control over their content or availability. When you have finished visiting any of these sites, you will be able to return to this page by clicking on your Web browser's "Back" button or icon until this page reappears.
- Teledesic
- Animated visualizations of an early version of the Teledesic network with 840 satellites
- SkyStation
- Motorola Space and Systems Technology Group and news release on the Iridium launch
- Alcatel
- Qualcomm
- Inmarsat
- TRW
- Teleglobe Canada
- Metricom
- Report on the wireless project in Mongolia
Copyright 1997 The New York Times Company
