Winner: Sprint's Broadband Gamble

Telecommunications

The number of traditional telephone lines is in sharp decline, and yet people are spending more on telecommunications than ever. The reason? Their money is paying for their wireless and high-speed Internet access. The logical culmination of these trends is a single mobile broadband service that would serve your every communication need—voice and data everywhere, in your kitchen, in your car, on the beach.

The idea hasn’t escaped some of the brightest minds in telecommunications. Two companies are now racing to unveil large commercial wireless broadband networks. Of the two, the one by U.S. long-distance giant Sprint Nextel of Reston, Va., is clearly in the lead: it expects to begin commercial service in three U.S. cities in April and to cover 70 million people by year’s end.

But Sprint won’t be the only game in town. Clearwire, a four-year-old company based in Kirkland, Wash., run by cellular visionary Craig McCaw, is aggressively converting an existing network to a mobile broadband one. The company hopes to cover another 30 million people and begin operating by mid-2008.

Negotiations that would have led to a broad operating agreement between Sprint and Clearwire broke down last fall, but the two still expect to allow roaming between their networks. Doing so will be easy; they both use the same technology, WiMax, based on the IEEE 802.16 wireless standard, in the same spectral band, 2.5 gigahertz.

The Sprint and Clearwire networks are just the sort that some other bright minds in the industry will need as they design the next generation of Internet-based communications devices. Google’s so-called gPhone is just the best known of the many innovations to come.

Sprint's service, called Xohm (rhymes with ”home”), will have its own operating company and brand. It will offer data rates between 2 and 4 megabits per second—better than some of the DSL and cable available in the United States. By contrast, Apple’s iPhone users are stuck in the slow lane at about 50 ­kilobits per second. Even the latest third-generation wireless services typically run at less than 500 kb/s, less than a quarter of Xohm’s speed. By blanketing entire metropolitan areas, and eventually the whole country, Xohm’s mobility and ubiquity will distinguish it from that other wireless broadband standard, Wi-Fi, which is mainly used for stationary, short range, indoor connections.

If you’re a certain kind of person (you probably are if you’re reading this magazine), the idea of having a Wi-Fi–like connection to the Internet everywhere you go is exhilarating. Armed with a Xohm-compatible smart phone, a networking card for your laptop, or a tiny ultra­mobile personal computer, you could stream movies or music in real time, and you’d be able to send video and sound files too—right from your kid’s soccer game, for example. It’s closer than you think: Samsung, of Seoul, South Korea, already has a Xohm-ready ultra­mobile PC on the market, and others are expected to follow. As long as you’re in a metropolitan area that gets the Xohm service, you’ll be instantly and always online at high speed.

Sprint, a long-distance company that is also the third-largest wireless carrier in the United States, is not turning its back on telephony. But the service is best thought of as a broadband network that lets you make voice calls instead of the other way around, as with just about every other cellular service in the world.

Sprint spent several years testing a number of different communications technologies before choosing WiMax. Xohm’s key equipment suppliers—Intel, Motorola, Nokia, and Samsung—are just a few of the many that manufacture WiMax equipment. And dozens of WiMax networks are in the works, including ones in Brazil, Ireland, and Japan. A network has been up and running in the Dominican Republic for several months.

Still, Sprint isn’t betting the farm on Xohm. It will continue to sell the phones and the regular voice and data services of its existing cellular networks, which accounted for all but US $1.6 billion of the $10.3 billion in revenues the company earned last year.

But Xohm is a high-stakes ­gamble nonetheless. Sprint is investing $3 ­billion in the new network. It has committed a wide swath of expensive radio spectrum. And by offering a data-­centric ­networking service, it is undoubtedly forgoing some revenue from voice calls, as well as from text messaging, ­ringtones, answer tones—just about everything that today makes a conventional cellular service profitable.

”Those are relevant to the traditional voice telecom companies, but we’re ­moving to a completely different model,” says Rebecca Hanson, Xohm’s vice president of strategy. ”We’re creating a whole new business.”

According to Jeffrey Davis, a telecom analyst at the Yankee Group, Xohm’s data rates will be high enough to motivate many to discontinue their home broadband service entirely, just as millions of regular cellular voice customers have dropped their wire-line phone service. Sprint has yet to set prices, but they are expected to be comparable to landline broadband rates, which in the United States start at about $40 per month.

Perhaps Xohm’s boldest gambit will be charging customers full price for the phones and other devices they’ll use on the network—unlike the heavy discounts that have become standard in U.S. cellphone deals. In return, however, Sprint won’t require two-year contracts, or any contracts at all. Users will be able to go month by month or even pay by the day or hour, like Wi-Fi users at an airport. You might prepay for a gigabyte of data transfers, for example, or buy a coupon that lets you upload 100 photos to your Facebook page. In fact, if Xohm is as successful as Sprint hopes, camera manufacturers—and plenty of other consumer device makers—will incorporate chip sets compatible with Xohm in their products, and you’ll be able to upload those photos directly from your camera.

Eventually, Sprint officials hope to lure entire industries. For example, gas and electric companies typically build large-scale communications networks to monitor their ­sprawling distribution networks. According to Warren Causey, vice president of Sierra Energy Group, Xohm ”offers ­utilities the ­advantage of no longer having to manage complex wireless networks.”

Perhaps most intriguing of all, Xohm will definitely roil the debate, currently raging among telecommunications executives and Internet companies, about what is called network neutrality. At issue is how much control carriers will exert over the networks that we use to access the Internet. The worst possible scenario has broadband carriers like AT&T and Verizon favoring one service over another, such as Yahoo over AOL, to such an extent that the Internet in effect fractures into pieces. It would become difficult or impossible to communicate with everyone. Even in such a case, a new, neutral nationwide broadband service—which Xohm aspires to be—could provide a haven from which the entire Net would be accessible.

Why WiMax? At least six different standards have been floated for mobile high-speed wireless. To understand why Sprint went with WiMax for Xohm, it helps to understand where the wireless network is today, and where it’s going.

Cellular carriers are at a ­crossroads. The convergence of the Internet and telecommunications—and in particular the carriers’ transition to the Internet Protocol from older technologies designed to carry voice—is occurring very slowly for traditional wire-line networks. But analysts are generally agreed that within a decade, wireless companies will have given up these older protocols in favor of breaking up voice calls into IP packets before sending them across high-speed networks.

If the wireless future will bring a consolidation of voice and data, it may also unify the communications protocols that underlie most of today’s cellular networks. Today we have two main cellular standards—GSM, by far the more popular worldwide, and CDMA, which is used by carriers in a handful of countries, including Sprint and Verizon in the United States.

In addition to these two voice-telephony standards (and a bunch of minor ones), there are several others for data, including (for GSM networks), Enhanced Data rates for GSM Evolution, the second-generation standard that iPhone users are stuck with, and High-Speed Packet Access (HSPA), a third-generation one mainly found in Asia and Europe.

WiMax, with its high data rates and reliance on IP, is generally regarded as a fourth-generation network. And by 2010 or so, it will face strong competition: a 4G cellular broadband successor to today’s GSM/HSPA networks, called Long Term Evolution (LTE), should be available, with data rates at least as high as WiMax’s current ones. Some CDMA carriers may also upgrade to LTE. There will be other contenders for 4G supremacy—Qualcomm, of San Diego, the inventor of CDMA, is pushing something called Ultra Mobile Broadband—but the main rivals will be LTE and WiMax.

The same basic transmission scheme underlies LTE and WiMax: Orthogonal Frequency Division Multiple Access, which is essentially what is used in the Wiâ¿¿Fi (IEEE 802.11) standard as well. OFDMA has, in other words, pretty much won the field for high-speed wireless data.

If OFDMA is the new standard, the road to today’s WiMax has been an unusual one in other ways. The core wireless technology, embodied in the first version of the IEEE 802.16 standard, harks back to the tail end of the dot-com heyday. It was designed for an immobile device that would sit in a home much like a DSL modem—except without needing an expensive wire or cable connecting it to the outside world. By 2004, wireless equipment makers were producing antennas and end-user devices that could exchange data at DSL-like speeds.

At the same time, though, one researcher, Hui Liu, now an associate professor at the University of Washington in Seattle, was solving the basic problems of mobile communications between OFDMA-based transceivers, such as handoffs from one cell tower to another. First at one company, Cwill Technologies in Austin, Texas, and then at another, Navini Networks in Richardson, Texas, a leading maker of WiMax equipment, Liu applied his work to the then-new fixed 802.16 standard. In early 2005, the South Korean government pushed the country’s leading telecommunications companies, such as Samsung, to commercialize mobile WiMax within 18 months.

What the Experts Say

Researchers elsewhere were skeptical, but by 2006, WiBro (as this South Korean version of WiMax was called) ­antennas were going up in Seoul and other cities. The version of the 802.16 standard that Xohm adheres to is a little different—it uses 10-megahertz channels, for example, while WiBro’s are 8.75 MHz, and a number of additions have been made, including several that fall under the rubric of ”quality of service”—but Korean eagerness to combine mobility and broadband speed can be credited with its breakneck development and much of the lead today’s WiMax has over LTE.

Although both LTE and WiMax use OFDMA, there are some differences of course, and according to Doug Smith, Sprint’s chief technical operations officer, the net result is that WiMax will be more spectrally efficient, packing more data transmission into the same bandwidth. For one thing, LTE uses a frequency-­division scheme: it divides its assigned spectrum into two different channels, one for transmission and the other for reception, with a guard band between them to prevent interference. WiMax uses a single channel and divides it by time—­sending, then receiving, then sending again. Frequency division is simpler, but the guard band wastes a small but precious amount of spectrum. In addition, the relative sizes of the two channels are fixed, while in a time-division scheme more time can be allocated as needed for sending or receiving.

Smith says that initially, the network will allocate about one-third of its bandwidth to uploading, that is, to letting users transmit files. That’s far more than other wireless—or wired—services do. ”If people are doing more downloading, then we can give even more time to it. We’ll have to see how the network is used. A video blogger might upload more than download,” he notes.

No one can say today just what LTE data rates will be, but we do know that WiMax does a lot better than HSPA. For a given 10 MHz, HSPA has a maximum throughput of 7.2 Mb/s downloading and 2.2 Mb/s uploading. WiMax, by contrast, gets up to 37 and 6 Mb/s, respectively.

In addition, Sprint has more of those 10-MHz segments than its competitors. It started to acquire spectrum in the 2.5â¿¿GHz band back in the heady dotâ¿¿com days of the late 1990s. Coincidentally, Nextel did as well, a key reason Sprint regarded it as an attractive takeover target. It has also been swapping snippets of spectrum with Clearwire in each other’s cities. All told, Xohm will have 120 MHz of spectrum in most of its markets. By comparison, AT&T and Verizon each only have about 50 MHz for their 3G networks today (which use the 1.9-GHz band).

Ideally, carriers like to divide each cell into three sectors, using directional antennas that limit themselves to 120 degrees of coverage. Such a scheme is called a reuse of 3. By using a different frequency in an adjacent cell, interference at the edges, where the cells meet, is minimized.

Ali Tabassi, Sprint’s vice president of technology, says that ”in order to truly deliver mobile wireless broadband, one needs to have 10 MHz each for the downlink and the uplink, with a reuse of 3.” A time-division transmission scheme like HSPA would need 60 MHz of spectrum to do that. That’s a little more than what AT&T and Verizon have, but it’s only half of Sprint’s holdings.

In the long run, LTE will have much higher data rates than HSPA. And because LTE is designed to be an easy upgrade from GSM and HSPA, it will let carriers get those higher data rates in their existing 1.9-GHz spectrum. Most GSM operators will move to LTE, so it’s also likely to have a broad base of suppliers, carriers, and users—an ecosystem—from the outset. On the other hand, Xohm, Clearwire, and other WiMax-based networks will have a three-year head start in the marketplace.

Building a new network from scratch is a gargantuan task—one that, fortunately, Sprint doesn’t have to undertake. For one thing, between its CDMA and other networks, the company starts with over 50 000 cell sites. Over the next year and a half, it will add WiMax radios and antennas to about 12 000 of them and create about 3000 new cell sites.

Smith insists that Xohm will be the fastest-built network in cellular history, and it just might be. Sometime this month Chicago and the Baltimore–Washington, D.C., area will get the service on a precommercial basis—a beta test of sorts. Full commercial service in those metropolitan areas is slated for this coming April. Other locales will quickly be brought online, culminating with New York City late this year. In broad terms, Sprint is working on tier 1 cities—the largest ones—while Clearwire concentrates on tier 2, places like Portland, Ore., and Grand Rapids, Mich.

Whether Xohm thrives will depend on more than just the number of people it’s available to. Equally important are the services it can conjure up—including the mysterious, as-yet-undefined ones that Hanson spoke of. Of course, would-be providers of these services will want to see the millions of users first.

And it’s actually a lot more complicated than that. The service providers—for example, online game companies that want to build mobile versions of their games—won’t take the plunge unless they see a panoply of devices—laptop cards, smart phones, ultramobile PCs, and so on—with which users can access the network. Yet hardware manufacturers are usually reluctant to design and build those devices unless there is a big, obvious market of customers—the same customers Sprint needs to attract in the first place.

Sprint plans to break into these vicious circles within circles by assembling its own WiMax ecosystem. It has chosen three big manufacturers—Motorola, Nokia, and Samsung—to make handsets and laptop cards, as well as the radios and antennas that will go on rooftops and towers, and the routers and servers that will connect them to the Internet as a whole. Sprint has gone to extraordinary lengths to ensure that the three vendors—and a fourth, Intel—will be able to test their equipment and its compatibility with that of other companies, including cellphone makers ZyXel of Anaheim, Calif., and ZTE in Shenzhen, China. Each of the four has its own area, which it can keep under lock and key, within Xohm’s own testing lab at its Herndon, Va., headquarters.

Perhaps the best thing that Sprint has going for it is that Xohm is merely a focal point for a growing interest in WiMax. In October, for example, Cisco Systems, of San Jose, Calif., the router giant, bought Navini Networks. In all, the WiMax Forum claims about 500 members. In addition to Cisco, some key ones outside the immediate Sprint orbit are Fujitsu, General Dynamics, Nortel Networks, Siemens, and Alcatel-Lucent, which was shut out of Xohm but built the system for the Dominican Republic that was turned on in October. Alcatel-Lucent has contracts to build WiMax networks in Europe, Brazil, China, and elsewhere. Meanwhile, phones using Google’s gPhone design may show up in the second half of the year.

As for the ecosystem of service providers, Xohm has several things going for it. First, as it is an open Internet platform, thousands of existing applications and services—Google search, instant messaging, iTunes, peer-to-peer music file sharing, you name it—will work just as they do with any other broadband connection. And according to Hwan Woo Chung, a vice president in Samsung’s telecommunication systems division, the network will not only be open enough but also fast enough to allow Internet-based television and other video transmissions to flow without noticeable latency.

Sprint will probably offer voice calling on the network, but users can bring their own Internet-telephony programs—Vonage, Skype, and the others—to Xohm. These programs will work as they do on landline broadband—generally pretty well, occasionally less so. The quality-of-service provisions in the WiMax protocol let a carrier protect certain types of transmissions, giving their data packets a higher priority as they travel through the network. Sprint would of course protect its own voice packets. Doing so raises, again, the question of network neutrality, because Vonage’s unprotected voice-over-IP service, for example, won’t work as well as Sprint’s own.

Sprint’s Hanson knows this is an issue. ”We’re not in the business of picking winners and losers,” she says firmly. ”We’re an open network. The more data transmitted across it, the more money we make, in the long run.” And indeed, a company like Vonage, of Holmdel, N.J., could pay Sprint to protect its packets equally. If Sprint is willing to do that for any voice-over-IP provider, and similarly treats other applications and application providers, then the network will, in practical terms, be neutral.

Net neutrality will become a bigger issue only as the Internet continues to mature and as the major carriers and service providers get larger in size and fewer in number. Two months ago, AT&T announced it would use a software program to inspect the YouTube videos that its subscribers try to watch, blocking any that might infringe the copyrights of the major record labels and movie and television studios. It’s just one example of how your Internet experience will differ depending on the company from which you access it—an early fracture, critics say, in the unity of the Net. Against that, a new Internet provider promising an open network—and a fully mobile, wireless one at that—will be a welcome addition.