Mobile Satellite Services

The ITU has been trying to set a global wireless networking standard for ten years. Although the benefits would be great-world travelers could use the same terminal everywhere, while equipment manufacturers would have to compete in an open market without vendor lock-in-they haven't succeeded.

Technology advances have made it possible to build "super GEO" satellites, so powerful that a device little larger than a cell phone can pick up the signal.

Depending on whose estimates you believe, cell phones can't reach between 20 percent and 50 percent of the United States. Though the uncovered areas account for only 1-5 percent of the U.S. population, that's still millions of relatively wealthy potential customers.

Many vendors don't want to adopt a system developed by a competitor, while governments are often unwilling to adapt their licensing regulations to match those of other countries. The closest the world has to a standard for cell phones is the Global System for Mobile Communications (GSM), but that's really five different systems, because of the different frequency bands it must work in. Likewise, the rival cdmaOne system is really two systems. And the incompatibility is set to last into third generation (3G) mobile, thanks to the same commercial and political rivalries, as well as the practical difficulties of upgrading older systems.

But 3G could still bring about mobile networks that work everywhere. In addition to the broadband networks that almost every cellular operator is now planning or building, 3G also includes an often-overlooked satellite component. Known as Mobile Satellite Service (MSS), it's intended to take over when a user is out of range of terrestrial base stations. Most MSS networks are designed to cover the entire planet, so they require that rival companies and governments set aside their differences and adopt the same technology worldwide.

MSS won't provide the high data rates hyped by cellular 3G, or even the more mediocre speeds already achieved in Japan and Korea. Its target is 100Kbits/sec, compared to the 2Mbits/sec that developers of Earth-based technologies claim. Actual performance, of course, will be lower. MSS networks must also overcome financial problems. Eight MSS providers were licensed a year ago, but some are already bankrupt, and it's unlikely that they'll all manage to get off the ground.

Instead of purely space-based networks, some MSS providers now want to augment their satellite capacity with repeater stations on the ground. Regular cellular service providers are understandably upset about this, though many are also building some satellite capability into their own networks in the form of Global Positioning System (GPS) receivers.

A GPS receiver isn't a true satellite phone-it can't transmit a signal back up-but the most advanced location-tracking systems do depend on a convergence between terrestrial and orbital networks that demonstrates how the two can work together to provide wireless connectivity everywhere. Combined cellular and satellite systems already allow people in poorer countries to make their first phone calls. The 3G versions might also allow them to send their first e-mails, as well as enable business travelers to move across America or the world without disconnecting from the corporate network.

THREE WAYS TO 3G

Most 3G systems currently under construction were developed as part of International Mobile Telecommunications 2000 (IMT-2000), an ITU project that has so far run for a decade. IMT-2000 is also the name of the official worldwide 3G standard developed by the project, though it can't really be described as such. The "standard" has so many mutually exclusive options that all the 3G systems launched so far can claim compliance with it, yet none actually interconnect with each other.

Though this sounds bad, today's IMT-2000 is actually a streamlined version that encompasses less than originally planned. Back in 1992, it was supposed to cover almost everything wireless. In addition to IMT-2000 cellular networks, there would be IMT-2000 Wireless LANs (WLANs), IMT-2000 satellite networks, and even IMT-2000 fixed wireless links. All would use frequencies in the same 2,000MHz range and ATM-based protocols.

The rationale behind this plan was twofold. First, basing all types of wireless hardware on a single radio system means that many people would need to buy only one device. Customers could use a home or office cordless phone as a cell or satellite phone. A satellite dish could double as a wireless local loop terminal. Second, equipment manufacturers would also save money, because components developed for one type of technology could be reused for another.

The fixed and WLAN components were abandoned within a few years. Fixed wireless systems work best at much higher frequencies, where there's more spectrum available and less signal dispersion along a line of sight. WLANs could work at IMT-2000 frequencies, and Europe's Universal Mobile Telecommunication System (UMTS) has the option of using a WLAN-like architecture to provide higher data rates inside a building. (This is where the hyped UMTS speed of 2Mbits/sec comes from; the outdoor data rate is only 384Kbits/sec.) However, there's no point wasting scarce licensed bandwidth on a 2Mbit/sec WLAN, when networks using free unlicensed spectrum are already more than 50 times faster.

This left IMT-2000 embracing only cellular and satellite. The ITU duly recommended two sets of frequencies that all countries were supposed to allocate to 3G: one for terrestrial and one for MSS. While many national regulators ignored the recommendation on terrestrial spectrum, almost all followed the one for MSS, making it the only IMT-2000 system that can be used worldwide. Most importantly, while the FCC chose not to release any new spectrum for terrestrial networks-U.S. cellular operators must deploy 3G in the same spectrum as their existing networks-it did do so for satellites.

After a long consultation process, the FCC finally licensed eight MSS networks in July 2001 (see table). With the exception of Boeing, which plans a network dedicated to air traffic control and in-flight entertainment, all hope to offer similar services to those of terrestrial 2.5G and 3G networks: data at around ISDN speeds, along with toll-quality voice, video, and multimedia messaging.

Six of the new systems are intended to cover the whole world, using a complex architecture similar to the existing Iridium and Globalstar networks. Because these systems require many satellites and billions of dollars, other companies plan to return to the single-satellite configuration that nearly all systems used until about five years ago. These can seem less exciting, but they have a greater chance of being constructed.

ROCKET SCIENCE 101

Most communications satellites are still placed in a high Geostationary Earth Orbit (GEO), meaning they appear to hang at a fixed position in the sky. This works well for fixed systems such as Very Small Aperture Terminals (VSATs) and TV tuners, which can simply point a receiver towards the satellite. It hasn't traditionally been as useful for mobile devices because the orbit is so high that signals are hard to pick up by the time they reach Earth. They require a large receiver, preferably dish-shaped and directed along a line of sight.

To provide a stronger signal, most designers of mobile satellite networks instead choose a Low Earth Orbit (LEO). This means that the transmitter is only a few hundred kilometers away, allowing a handheld phone to pick up the signal. The phone doesn't have to be pointed in any particular direction, but it does need a line of sight to the satellite, so it will only work outside.

About twenty separate consortia proposed LEO constellations in the 1980s and 1990s, known collectively as Global Mobile Personal Communications by Satellite (GMPCS). These systems differ from MSS in that they use higher frequencies than 3G cell phones, and don't offer the same advanced 3G services.

Of the twenty GMPCS constellations, only three have actually been built in LEO orbits. Iridium and Globalstar both offer voice and narrowband data, at respective speeds of only 1.2Kbits/sec and 9.6Kbits/ sec. Both claim to be faster, but this is using compression, which won't work with applications such as VPNs. Orbcomm offers messaging, but with a high latency (up to whole minutes) as users sometimes have to wait for a satellite to come into range.

Though technically sophisticated, none of these networks has been commercially successful. The builders of all three have at some point filed for Chapter 11 protection, a trend that predates the current telecom and dotcom recession. The problem is competition from terrestrial cellular networks, which already reach nearly every lucrative market. Most people in the world are still not covered by cellular, but that's because they can't afford to buy or run a cell phone, let alone a satellite phone.

Other would-be GMPCS providers have cut back their plans. Bill Gates's Teledesic was once the most ambitious, claiming that it would have an 840-satellite system operational by now. Instead, it has repeatedly scaled back its plans, first to 288 and now to only 30. Teledesic has also been delayed, now promising global coverage by 2005, and emphasizing fixed, rather than mobile, services. The ITU says that it must launch at least one satellite and sign up some customers by September 2004, or lose its spectrum.

These delays and financial problems also affect the new MSS licensees. The most ambitious of them on paper is Constellation Communications, whose phone has been disconnected and its Web site taken over by a domain name reseller. Ico Services is the only one that's yet launched a satellite, and it's already been through bankruptcy.

This doesn't mean that MSS won't happen. Ico emerged successfully from bankruptcy, as did Iridium, and both continue to launch new satellites. Iridium even plans to upgrade its network to MSS capability. Globalstar says it will do the same, though perhaps not using LEOs. Technology advances have made it possible to build "super GEO" satellites, so powerful that a device little larger than a cell phone can pick up the signal.

Globalstar could serve the whole planet using only four GEO satellites, and two other MSS licensees plan to serve North and South America using one each. These two won't be as useful for world travelers, but they could help people in the less-populated areas of the United States. Depending on whose estimates you believe, cell phones can't reach between 20 percent and 50 percent of the country's land area. Though the uncovered areas account for only 1-5 percent of the U.S. population, that's still millions of relatively wealthy potential customers.

Three other companies are already providing GMPCS services via GEO satellites, though none offers a cell phone that will work in the United States. Thuraya offers voice and 9.6Kbit/sec data throughout the Middle East, Africa and Europe. Asia Cellular Satellite (ACeS) serves Asia, with an ambitious goal to bring telephony and 2.4Kbit/sec data to 50,000 rural villages within a year. A single GEO satellite is cheaper than a LEO constellation, so it says it can do so at a price these villages can afford.

Both Thuraya and ACeS are based on GSM, making dual-mode phones easier to manufacturer, though they don't provide all the advanced features GSM users enjoy. The Short Message Service (SMS), for example, only works when connected to a terrestrial network.

Inmarsat covers most of the world, at data rates up to 64Kbits/sec. (see "Broadband Reaches Afghanistan," January 2002). Unfortunately, the terminals for such speeds aren't yet down to cell phone size-they're closer to a laptop, described as "portable" rather than mobile. Inmarsat does plan a broadband GMPCS network, which will support voice and smaller terminals within two years.

EXTRA TERRESTRIAL

MSS providers have a strong incentive to build their networks. The FCC has said that they'll lose their spectrum licenses unless they meet several deadlines, including full commercial service by 2007 (see Space Timetable). What will happen to any spectrum reclaimed by the FCC is the subject of fierce debate.

Cellular operators have already campaigned for the extra spectrum to be opened up to terrestrial 3G, but so far regulators have resisted the lure of auction revenues: One reason the FCC took so long to issue MSS licenses is that it spent many years examining (and rejecting) arguments from cellular operators. Regulators are bound by international treaties, and though these might be torn up if the satellites aren't actually launched, the FCC has further concerns about universal service. North America is larger and less densely populated than Europe, making satellite the only way to cover the entire U.S.

Some MSS providers have an alternative, though one that upsets the cellular operators even more. Ico still plans to build all ten of its satellites, but has also filed an application for an "ancillary terrestrial component": a series of base stations in major cities that would provide a much stronger signal than that from the satellites. This would enable MSS phones to work inside, as well as in "canyon" streets where buildings block a phone's view of a satellite. ACeS and Inmarsat already do this in remote areas of Africa, using base stations mounted on trucks to roll out cellular networks where no infrastructure exists.

There's a big difference between building a temporary network in an area with no existing telecom services and building a permanent network in major U.S. cities. Cellular operators say that these amount to cellular networks, using valuable 3G frequencies: Some of them have paid billions of dollars for 3G licenses in a single country, so they don't like the idea of competitors getting a worldwide spectrum allocation for free.

Even if MSS operators don't get to build repeater stations, their users should still be able to roam on to regular 3G networks, though Ico says that that this would be an inferior experience to using the same provider for both types of service. And even without MSS, 3G is deeply entwined with satellites.

There still aren't many true 3G phones on sale or networks operational, but most of the prototypes so far incorporate GPS receivers somewhere-in either the phones, base stations, or both. The most advanced location technologies effectively use cellular base stations as GPS repeaters, enabling the signal to be picked up by users deep in a building, underground, or otherwise unable to receive the satellites' own relatively weak signal.

This repeater function is, at least at first, the main added value that 3G location technology will offer to GPS. Cellular operators plan to offer downloadable content tailored to suit a particular location, but present networks aren't up to the task.

"There isn't enough bandwidth to stream the map," says Shay David, vice president in charge of development at Powerloc Technologies, a company that makes navigation software. "We'll get to streaming eventually, but the hardware today is not ready for it." A modern PDA can store a detailed map of the entire United States on a Compact Flash card.

The fastest 2.5G or 3G networks available so far in the United States run at an absolute maximum speed of 144Kbits/sec, with real speeds much less than that. Even accepting the hyped data rate, each megabyte takes almost a minute to transfer, and detailed maps can run to many megabytes. Other countries have faster networks, but usually with prices to match.

Whereas Verizon charges only $100 per month for unlimited access at dial-up speeds (roughly), the per-packet charges on NTT DoCoMo's popular i-mode service mount up to more than $30 per megabyte. Costs should drop as faster networks come online, but physical limits on spectrum mean that mobile broadband might never reach the masses through cellular networks alone. Streaming maps on demand over satellite would be prohibitively expensive for all but life-and-death situations.

The best solution to this and most other wireless problems is a combination of satellite, cellular and WLAN connectivity, just as the ITU originally planned. However, such a combination would require a degree of cooperation among service providers that's currently lacking.

Senior editor Andy Dornan's book, The Essential Guide to Wireless Communications Applications, ISBN 013-0097-187, is published by Prentice Hall. You can reach him at adornan@cmp.com.

Space Timetable

With other satellite networks taking years or even decades longer to build than intended, the FCC is holding all eight Mobile Satellite Service (MSS) licensees to strict deadlines. If any company fails to meet all of them, it automatically loses its license.

Mobile Satellite Services