Are changes in the air for local access?(wireless broadband access)

Despite a slow start, wireless broadband access could be poised to make a breakthrough in '99.

When FCC Commissioner Michael Powell spoke to a conference of competitive local exchange carriers (CLECs) late last year, he urged them to build their own infrastructure wherever feasible, rather than relying on incumbent LECs for access. For carriers committed to wireline last-mile access, that must have sounded like a bad joke: The obstacles to laying a new, parallel local network are so overwhelming that even huge competitors like AT&T and MCI have come up far short. Similarly, smaller metropolitan fiber-based competitors have had to settle for connections to a select few buildings in big cities.

But there are facilities-based CLECs putting in large swaths of last-mile infrastructure right now. They are basing their networks on wireless technologies, and although these new broadband providers are still struggling to build a customer base, there are signs that they could begin to go mainstream during 1999.

The two largest of the new wireless competitors, WinStar (www.winstar.com) and Teligent (www.teligent.com), have been marketing voice and data services to tenants of large buildings in areas where such buildings are clustered. The initial target market was limited by the wireless carriers' point-to-point technology, which demanded economies of scale similar to those that have limited fiber CLECs to serving the largest buildings.

This year, however, wireless broadband is poised to penetrate more deeply into the user base, because the providers are starting to migrate from point-to-point connections to point-to-multipoint systems, which can connect multiple buildings to a single base station, radically changing the economics of providing broadband wireless. Once that technology is in place, broadband wireless won't just be for the big sites any more.

The Promise of Broadband Wireless

On the face of it, broadband wireless seems ideal for solving the last-mile bottleneck. While new wireline providers must battle to get municipal rights-of-way, a wireless provider needs only to place a "hub" site within line-of-sight for multiple buildings, and then convince the building owners to allow an antenna on the roof [ILLUSTRATION FOR FIGURE 1 OMITTED]. Then, it's simply a matter of transmitting between the two sites, on licensed frequency bands ranging as high as to 38 GHz.

For providers, the cost comparison with wireline isn't even close. William J. Rouhana, Jr., chairman and chief executive officer of WinStar, estimates that at $20,000 per building, a wireless connection costs him less than one-tenth what it would cost to run fiber to the same site. And that's for point-to-point connections. As multipoint systems proliferate, the cost will fall to about $6,500 per building connection.

Wireless providers also promise that they can provision service in days rather than the weeks or months it can take traditional wireline carriers to get a service up and running. And the wireless systems support almost any kind of service a customer needs - local and long distance voice, dedicated Internet access, ATM, frame relay or traditional T-carrier service. Because the costs of building the network are so low, WinStar, Teligent and Advanced Radio Technology (ART; www.art-net.net) quote prices as much as 30 percent below the ILECs'.

The sales pitch, according to consultant Craig Mathias of Farpoint Communications Group, is, "'We're the new telephone company. And we don't want you to think of us any different from any other telephone company, other than we're a lot more responsive and we will save you money.' They all come in with that attitude. Rarely do they get into detailed discussions of technology."

Part of the reason they don't get into technology is that the target market - small to medium-sized companies - is less likely to have a technology guru looking to grill the sales rep. But there's another factor: The technology may not even end up being wireless. WinStar, for example, jumped into the access market by reselling ILEC services and infrastructure, opting not to wait until its own networks were finished before starting to acquire customers and experience in the business.

In fact, most of WinStar's lines still aren't provided over its wireless networks. According to the company's 3Q98 report, only "over 18 percent" of its lines are "on-net" - i.e., connected to the wireless network. The company points out, however, that in New York, its largest market and the one where it has operated the longest, over 54 percent of the lines are now on the wireless network. WinStar also claims that one-third of the lines in its five most mature markets are "on-net." Still, while WinStar wants to serve customers primarily from its own network, that's clearly going to take a little time.

To encourage more customers to come "on-net," as well as to position itself as a one-stop shop, late last year WinStar kicked off a promotion called "Project Millennium," which features up to a year of free local phone service for customers who sign up for broadband access from a building with a wireless connection. As a result, the company claims that in New York, 93 percent of its orders for November 1998 came from "on-net" buildings, though the totals for other cities range as low as 56 percent in Los Angeles.

For now at least, the wireless providers are small potatoes in local access. WinStar claims 12,000 customers - both wireline and wireless - while Teligent spokesman Robert Stewart concedes his company still has "less than a thousand" accounts. However, these new companies can grow much more quickly than their counterparts offering wireline competitive access; instead of digging a new trench to each new building they want to serve, their multipoint technology lets them open up service to entire city blocks with the installation of a single base station.

No matter how fast technology advances, back-hoes don't obey Moore's Law. Neither do ILEC legal departments. As a result, there are nearly insurmountable barriers to widespread facilities-based local competition. But the old rules don't apply to wireless. In fact, wireless is the only last-mile technology that really heeds the mantra of the Information Age: faster, better (or at least as good), cheaper.

Getting to the Point

The early incarnations of wireless broadband access networks featured point-to-point connections. Transceivers sat on the customer's site and at the provider's "hub," where the network interconnected with a fiber backbone [ILLUSTRATION FOR FIGURE 2 OMITTED]. That configuration remains the most cost effective for a service provider that needs to connect only a single building in a given area. But increasingly, providers want to set up point-to-multipoint configurations in which one base station can communicate with several buildings. In larger areas, several base stations may be needed, and these would be networked together along the fiber backbone [ILLUSTRATION FOR FIGURE 3 OMITTED].

WinStar and Teligent already claim to be delivering multipoint, but ART's chairman and CEO, Henry C. Hirsch, has been a little more cautious about deployment timetables. He said, "We're looking at another six to nine months before we have point-to-multipoint in commercially available quantities."

But once multipoint does roll out, the addressable base for wireless broadband will expand dramatically. "With point-to-point, you pretty much don't go below a 75,000-square-foot building, roughly," said Bill Maxwell, president and COO of ART. "But when you get down to point-to-multipoint, you could pick up 5,000- or-10,000-square foot buildings if they have one or two good-sized users in there."

WinStar's Bill Rouhana agrees, noting that if providers can serve 10,000-square-foot buildings, "That pretty much means everywhere, at least in the commercial marketplace. If there's line of sight from a hub to a building, [the wireless provider is] going to go to that building some day. It's just a matter of doing it in an orderly way."

In addition to expanding the number of customers wireless providers can reach, multipoint also offers more advanced networking technology than point-to-point's relatively dumb pipes, explained Cynthia Hillery, VP of marketing for Netro (www.netro-corp.com), a manufacturer of multipoint access infrastructure equipment. Wireless careers will use advanced protocols and ATM-based technology to statistically multiplex multipoint traffic at the hub. In turn, that will allow careers to offer dynamic bandwidth allocation and quality of service to their base of multipoint customers.

Reliability

Once the network infrastructure is built out, wireless broadband providers face another hurdle: convincing users that their service is every bit as reliable as landline. Service guarantees and general perceptions are crucial when it comes to wireless, particularly because wireless is perceived as less reliable than wireline. Consultant Craig Mathias acknowledges the doubters but thinks they're wrong. "There are people out there who are very skeptical as to whether or not wireless works. But I've been in this business a long time. It definitely works."

Mathias and others note that the limitations of wireless technology have a greater impact on carriers than on the ultimate customers. For example, the first limitation stems from a basic principle: The higher the frequency, the shorter the distance a signal can travel without degrading.

In practice, that means that higher-frequency systems need more cells and more base stations, hence have a higher cost to build, according to Teligent's Robert Stewart. That represents an advantage for Teligent, which uses 24-GHz spectrum, over many of the other providers, who have primarily 38-GHz allocations, Stewart claimed.

Climate also affects how expensive a network is to build. "The only thing that degrades signal strength on high-frequency microwave systems is very large, heavy raindrops...they absorb the energy," said Netro's Cynthia Hillery. "So you know the physics of the rain characteristics in the area and you engineer your power levels and distances accordingly. You blend the frequency, the distance and the characteristics of the geography to engineer to a certain level of reliability."

And the level of reliability can get very high. WinStar's Rouhana notes that, because broadband wireless access networks are so dramatically less expensive to build than wireline, the providers can spend extra money on redundancy, backup and top-of-the-line equipment. "We are more than happy to translate that into guarantees to our customers," he said. "We think ultimately what will distinguish us in the marketplace is going to be the superiority of our network and the way we treat our customers." He claims WinStar has a lower rate of troubles per hundred lines reported to the New York Public Service Commission than any wireline carrier operating in New York City.

Competition

There's a related availability issue that Rouhana believes works in his favor. He claims WinStar has already signed up 100 of the Fortune 1000, and that much of the appeal for large business customers is that having the WinStar antenna on a roof gives customers "true" route diversity.

"We come into the building in a different way than the wireline providers, who all come in through the basement" he explained. "They all come in through the same conduit, they all are subject to the same crazy flying backhoes driving down the streets of the cities. If they cut one line, they cut them all, and if there's a flood in the basement, they all go. If there's a fire in the basement, they all go. Whereas we're up on the roof. Most of the time, if the basement goes, the roof doesn't."

Rouhana's perspective on building entry points is sort of a metaphor for how he views the competitive landscape: His principal competition is the guys who come in via the basement, not the other newcomers vying with him for space on the roof. "When you're in my seat, what you're really thinking about is not what Teligent or ART are doing. You're thinking about what the ILEC is doing, how to take customers away from them and how to distinguish myself from them. They're who the customer knows."

One key differentiator may be service provisioning. While all the wireless providers tout their ability to fulfill orders more quickly than the ILECs, Rouhana says the issue isn't how long it takes to get a service up and running, but how reliable the provider is about meeting commitments. Rouhana says that he now promises 45-60-day provisioning time, "even though we can do it much faster. But the marketplace, in my view, doesn't require it.

"What is required, however, is delivering on what you promise," Rouhana continued. "The big problem most customers have is when someone promises something but doesn't deliver it on time. Then customers find themselves in trouble."

Conclusion

While off to a slow start, wireless broadband networks could pick up speed quickly. However, it's important to note that this service isn't the same thing as wireless local loop. For example, at just under $900, ART's T1 service is a formidable competitor for an equivalent wireline service, but no match for a DSL offering that could provide as much bandwidth, at least in the downstream direction, for a fraction of the cost. Plus, even with the emergence of multipoint technology, none of the major wireless broadband players show the slightest interest in going to the residential/telecommuter market.

Instead, they're targeting business users who want the high-speed access that has been prohibitively expensive up to now. To deliver that access, they're not committing exclusively to wireless delivery. Most say they'll resell ILEC lines or network elements if that's the best way to push service down the last mile to a given customer.

As consultant Craig Mathias puts it, "I think people really view this as just another alternative in terms of providing the high-capacity networks that they need."

Eric Krapf is BCR's managing editor.