What Will LMDS Hatch?

Local telcos and cable TV operators, this is your wake-up call. Although it has been unclear just how long the gestation period for local multipoint distribution service would be, the Federal Communications Commission is finalizing plans that could grant individual license holders up to 1.3 GHz of spectrum per market. That's 43 times the bandwidth obtained by winning bidders in the first broadband PCS auctions. And it represents more than twice the bandwidth of AM/FM radio, VHF/UHF television, and cellular telephone combined (Figure 1).

LMDS promises a wireless alternative to fiber and coax. Although the Bell regional holding companies and leading cable TV operators having been talking about, testing and promising various methods of delivering interactive broadband services for years, reality has not kept pace with the hype. LMDS may prove to be the shortest path to mass delivery of two-way video and high-speed data services.

Wes Vivian of Vivian Associates, an Ann Arbor, Mich.-based consultancy, says it should cost an LMDS operator less than $500,000 to put its first two-way cell-covering five to 10 square miles-into service. Additional cells require only repeater-hubs and will cost much less. In contrast, broadband fiber and coaxial cable networks must pass virtually every home and business in a given area before they can begin offering service. Even if the cost were the same, it would be almost impossible to build fiber to the curb at the same pace as wireless, Vivian points out. Thanks to its lower entry cost, LMDS is the lower-risk solution.

The ability to deliver true bandwidth-on-demand is another key LMDS advantage. With LMDS, any or all the bandwidth is available to all potential customers within range of a hub from the moment it is switched on. With coax or fiber, the last 100 feet still must be wired before a new customer can be activated.

There are some technical challenges associated with LMDS, including the cost of subscriber equipment. But the biggest hurdle may be psychological.

The radio industry has long been ruled by skeptics who believe there will always be an acute spectrum shortage. Like the British economist Thomas Malthus two centuries ago, they see only rising demand for fixed resources. Just as Malthus was blind to the gains of agricultural technology, spectrum skeptics fail to see that the migration to digital radio and higher frequencies will cause a temporary spectrum glut. LMDS is just one of several new allocations that will help overthrow the old paradigm that says radio is only good for narrowband services.

Fish or fowl? There is no consensus about what LMDS will be. Robert Shearing, chairman and chief executive officer of prospective LMDS bidder SkyOptics, San Diego, says that the market will vary tremendously between basic trading areas. "LMDS is no one thing," he says.

Industry participants see three major opportunities. LMDS pioneer CellularVision, New York, believes interactive television is the largest market. The firm is already using the 28 GHz band to deliver the equivalent of 49 channels of CATV service to parts of New York City that have resisted deployment of coax cable. CellularVision also plans to gradually introduce two-way services such as video-on-demand and high-speed Internet access starting at 550 kb/s. This will be accomplished by inserting uplink channels between the video channels using reverse polarization.

The second potential market is high-speed data. LMDS can not only provide access to the Internet, it can be used to interconnect local area networks, create campus area networks and serve as a medium for asynchronous transfer mode and Sonet. These capabilities permit LMDS operators to focus on business customers first-those best able to afford 28 GHz subscriber equipment before the development of a mass market. As Douglas Lockie, vice president of millimeter wave subsystem manufacturer Endgate Corp., Sunnyvale, Calif., quips, "It will be easier selling broadband wireless services to companies trying to interconnect LANs at reasonable prices than selling yet another cable TV service to holders of overdrawn credit cards."

How fast is LMDS? Hewlett-Packard suggests an asymmetrical system can operate at 1.5 Gb/s downstream and 200 Mb/s upstream. Even during peak hours, a heavily loaded asymmetrical system would yield 7 Mb/s downstream and 1 Mb/s upstream per household-enough to view two movies simultaneously.

Third, because it can handle thousands of voice channels, one should not discount LMDS' potential for providing voice service. It is unlikely LMDS will be built just for POTS-but voice could represent a good source of incremental business. Because mobile access is problematic at 28 GHz, though, LMDS is not likely to compete with cellular telephone and personal communication services (PCS).

There is considerable interest in LMDS-like offerings abroad. Countries lacking a modern telecommunications infrastructure see wireless as a means of leapfrogging into the 21st century. In Kobe, Japan, a 23 GHz broadband network was used to quickly restore CATV service after last year's massive earthquake. While the U.S. struggles with issues like band sharing and relocating incumbent users, Canada, Latin America and other areas contemplate allocating wide swathes of spectrum for broadband applications. The hatchling

LMDS solutions are just beginning to emerge. In July 1996, after a long battle with fixed satellite service vendors Teledesic and Hughes Space and Communications Co. (during which Teledesic described CellularVision's service as a "fifth way to get Gilligan's Island reruns") the FCC made some initial spectrum allocations for LMDS. Bandwidth of 850 MHz-between 27.5 and 28.35 MHz-was allocated on a primary basis, and 150 MHz-between 29.1 and 29.25 MHz-was allocated on a co-primary basis. The FCC also has proposed allocating the 300 MHz between 31 and 31.3 GHz. According to HP, the additional allocation would allow for greater separation between transmit and receive frequencies, eliminating the need for filters and ensuring efficient use of LMDS' spectrum.

In addition to CellularVision, which developed its own technology, both HP and Texas Instruments plan to offer LMDS equipment. LMDS manufacturers have to make decisions regarding cell size, how to divide the spectrum between transmit and receive functions, and antenna configuration-roof- or window-mount. But the key challenge for anyone interested in creating a mass market will be developing low-cost 28 GHz transceivers.

The transmitter amplifiers for a 28 GHz system require sophisticated semiconductor technology such as monolithic millimeter wave integrated circuits based on gallium arsenide. Until recently, the highest frequency band in wide consumer use was 2.4 GHz, found in microwave ovens. Within the last two years, direct broadcast satellite (DBS) has experienced rapid growth at 12 GHz, but the subscriber device is receive-only. Otherwise, products above 2.4 GHz serve industrial and governmental users.

There is less concern about the cost of LMDS infrastructure, which should compare favorably with that of cellular and PCS technology. Once an operator has constructed an LMDS switching center, additional cells can be constructed using fiber- or microwave-fed remote transceivers, or even using excess LMDS capacity. The typical LMDS hub will consist of a 27 GHz upconverter/downconverter unit, with most of the intelligence residing at the switching center.

Five major issues Five technical issues are hotly contested in the LMDS industry:

• One-way vs. two-way transmission. LMDS pioneer CellularVision entered the market with a one-way system, essentially a 27 GHz wireless CATV solution, and plans to offer two-way services as an upgrade. HP and TI plan to offer two-way systems from the start.

• Analog vs. digital systems. Unlike CellularVision, which offers an analog system, HP and TI plan to market higher-capacity digital systems. The main drawback to digital is that the vast majority of today's television sets, cameras and studio gear are analog. Digital transmission requires either an end-to-end digital solution such as high-definition TV, which tends to be expensive, or conversion back to analog at each end of the link.

• Reverse polarization. In theory, an antenna/receiver unit can listen to horizontally polarized signals from its local hub and ignore vertically polarized signals from an adjacent node on the same frequency. CellularVision and TI plan to employ reverse, or alternating, signal polarization in their LMDS equipment. However, HP Marketing Manager Gary LaBelle cautions that signals could become depolarized in some environments. And because the FCC has allocated all the LMDS band to a single licensee in each area, squeezing out more bandwidth may not be so important.

• Line-of-sight access. According to Endgate's Lockie, signals at 28 GHz propagate like "fat laser beams" and may be blocked by foliage. It is best to communicate above the tree line. HP plans to use line-of-sight communication from rooftop antennas to elevated hubs. It may be possible to overcome lack of line-of-sight with increased power, but that brings its own set of challenges, including possible higher transmitter costs.

If line-of-sight is an absolute requirement, operators will need to acquire hub sites at least 50 to 75 feet above the ground, which means LMDS will have a tougher time in residential than commercial areas. Business districts are easier to cover because there are more tall buildings and fewer trees. The hub antenna does not always have to be placed at the highest spot to reach all the buildings in an urban center. After all, the antenna only needs to "see" a piece of each building. And in commercial districts, one does not usually find the resistance to antenna towers encountered in residential areas. In some cases, though, customers located in leased office space do not have access to their building's roof.

In residential areas, using light poles may be an option but would require more cells and still would not guarantee line-of-sight to each home. TI believes line of sight is required to ensure link availability but that it has other virtues. Operation at low power levels requires overlapping cells, providing multiple communications paths.

• Extending coverage with reflections off large objects. CellularVision, which has 6500 subscribers in four cells covering Brooklyn and lower Manhattan, claims it is exploiting reflections off buildings to provide coverage. Endgate's Lockie, however, points out that reflections can cause multipath interference. In addition, reflections may be blocked by trucks or other large objects.

Efforts are underway by the Digital, Audio, Visual Council to create a global LMDS standard. According to Horen Chen, vice president of wireless broadband products for Stanford Telecom, Sunnyvale, Calif., the goal of the LMDS standards effort is to hasten the attainment of economies of scale. The standard addresses both the air interface and media access control layer but leaves issues such as reverse polarization and one-way transmission to the vendor's discretion.

Emerging questions The first question prospective operators-and equipment vendors-must answer is whether LMDS is likely to succeed first as a next generation cable TV service, or as local access for high-speed data. The answer will help determine the target price points, whether to focus first on residential or on business areas, and bidding strategies for the 487 basic trading areas.

CellularVision's strategy centers on the CATV market. In New York City, the firm's basic service sells for $29.95 a month. CellularVision claims infrastructure and operating costs per subscriber for its one-way analog network are around $300, in comparison with $700 for multichannel multipoint distribution service, operating in the 200 MHz range. LMDS also compares favorably with DBS, which CellularVision estimates at a cost of about $800 per subscriber. Unlike the other wireless cable solutions, LMDS is moving rapidly toward two-way transmission. And DBS has the further drawback of not offering local programming because each satellite beam covers a large geographic area (Table 1).

HP and TI look at LMDS somewhat differently, each planning to offer a digital system that can be used in either consumer or business applications. One of HP's top priorities is cost reduction-especially for the subscriber unit. In large quantities HP can get the cost of the subscriber transceiver/antenna unit down to about $350. Customers will still require either a radio modem for data at an additional $350 or a TV decoder for video at an extra $200. HP believes, however, that LMDS infrastructure will cost about $150 per customer covered, compared with $1000 for fiber to the home. At 25% market penetration, fiber would cost $4000 per customer while LMDS would cost $600 per customer.

TI's Communications & Electronic Systems has developed a two-way LMDS system called MulTIpoint. TI's strategy differs from HP's in that the firm aspires to be an LMDS system integrator. TI wants to offer everything from network access, video headends and telephony switches to set-top boxes and high-definition televisions. Like HP, TI is bullish on LMDS' low entry cost, claiming it will be half that of hybrid fiber/coax at 30% market penetration. TI also points out that a "baseline" asymmetrical configuration could provide 224 digital video channels and 16,000 telephone channels per residential node, or 192 T-1 circuits and up to 4608 DS-0 circuits per business node.

The challenge for LMDS operators will be creating deployment strategies. Unlike cellular telephone operators, which were able to start with large cells and gradually split them into smaller cells, digital LMDS operators must start with small cells. While broad coverage was the key to success in cellular, identifying and serving areas densely populated with potential customers is key to LMDS. Instead of splitting big cells, LMDS operators must craft a strategy based on adding small cells.

Not the same old species In LMDS auctions planned for late 1996 or early 1997, the FCC will probably give some advantages to small businesses with less than $40 million per year in revenue and less than $500 million in total assets. Possibilities include a smaller participation deposit, a 25% discount off the amount bid and low-interest, 10-year government loans. It is also likely the FCC will limit an RHC's participation in its own backyard, perhaps limiting it to 10% LMDS system ownership in any area in which it also provides local telephone service.

So far, however, very few firms have announced intentions to bid in LMDS auctions. It would seem likely the local exchange and interexchange carriers will be players. Among the Bell companies, Bell Atlantic, U S West and BellSouth have exhibited varying degrees of interest. Wes Vivian believes prospective bidders may be keeping silent because they do not want to do anything that might drive up auction prices. Instead, he points to cash raised by CellularVision in a initial public offering as an indicator of LMDS' perceived value.

Potential bidders should note that LMDS rules do not tell carriers what to do with the spectrum. It is possible that some players will acquire LMDS licenses with the goal of becoming wireless broadband service resellers-similar to what NextWave Telecom Inc. is attempting in broadband PCS.

It may take a while for people to accept the notion that LMDS isn't the same old species of narrowband radio. Users want more bandwidth, and they want it on demand. LMDS promises both.

Ira Brodsky is President of Datacomm Research Co., Wilmette, Ill.