That's because high-speed technologies like digital subscriber lines (DSL) and cable modems found in major cities are nowhere to be found in this town of 12,000. Instead, the data is being pushed over a fixed wireless connection using a 140-foot-tall steel tower just on the outskirts of town.

Atop the tower are circular antennas that are the workhorses delivering Internet access to this small town and other rural regions of the U.S., where fixed wireless is taking hold and going head-to-head with DSL and cable.

WILDFIRE GROWTH The investment research group Frost & Sullivan expects the broadband fixed wireless equipment market to grow at a compounded annual rate of 70 percent over the next four years. And other research firms are also predicting it will be a big source of revenue.

The Strategis Group projects broadband fixed wireless revenues will climb from $11.2 million in 1999 to approximately $3.4 billion in 2003. Much of the growth of fixed wireless in the U.S. can be attributed to communities like Keokuk that need a rural solution for broadband delivery.

But fixed wireless is not without its shortcomings. The solution, which uses radio frequency spectrum to transmit data from central offices to antennas in the home, still has major hurdles to leap over. To get up and running, a fixed wireless company needs the co”peration of local government, favorable topography, and almost unlimited access for antennas and equipment.

FERTILE GROUND Fixed wireless wasn't the first proposed "last mile" technology for Keokuk. Initially, DSL providers promised high-speed connectivity -- but failed to deliver. And because of its sparse population, cable access providers stayed away from the area. That's when Interlink, the major Internet service provider (ISP) in Keokuk, decided to get involved in the broadband equation.

What Interlink brought to the table was its strong track record making fixed wireless work in several of the other towns and counties it serviced. And a bargain of sorts was crafted between the two parties. In return for high-speed Internet access, the town gave Interlink something it needed -- total access to residents.

A similar success story with fixed wireless is echoed thousands of miles away in Sioux Falls, South Dakota. It is here that Sioux Valley Wireless found a community that previously had low data rate dial-up modem access to the Internet. Through a fixed wireless solution, SVW provided high-speed access to the rural area.

OBSTACLE COURSE One of the big stumbling blocks with fixed wireless is trying to secure placement of antennas. Stephen Webber, president of Interlink, says his company forged a relationship with the town and county agencies that helped it overcome this obstacle, so it could place antennas on the top of water towers and county buildings.

Another potential obstacle is physical: the topography of the land. In the case of Keokuk, the landscape and terrain is flat, so the line of sight needed between antennae is possible, but this is not so in all locations. A metaphorical obstacle is that potential customers get nicked for an installation charge when the technology is first rolled out.

Roland Van der Meer, a general partner at ComVentures, a Palo Alto, California-based venture capital firm that invests exclusively in communication technologies, says fixed wireless is a strong rural play where there are no established cable lines, but the technology must now compete with satellites for the rural last mile.

Acromegaly, or "giantism," starts at an early age. It is a life-threatening disease in which the body overproduces human growth hormones, causing bones and organs to thicken and become unusually large. It affects a relatively small portion of the population -- a total of 40,000 people in the United States, Europe, and Japan.

There are some effective drug treatments, but rare diseases like acromegaly don't get the attention of drug companies. Because the patient population is so small and the cost of developing drugs is so great, medications to treat these diseases haven't been pursued vigorously.

These kinds of diseases, which afflict fewer than 200,000 U.S. patients a year, are known as "orphan diseases," (for more on orphan diseases, see "Adopting orphan diseases," January 16). But the more than 6,000 rare disorders, taken together, affect approximately 25 million Americans, according to the National Organization for Rare Disorders (NORD), a federation of health organizations dedicated to helping people who have such diseases.

Scientists believe the information obtained from the Human Genome Project (HGP), the $3 billion effort to map our genes, will reduce the time and cost needed to find cures for rare diseases. This could open the way for startup drug companies to pursue these diseases.

As part of our series of articles on the radical changes facing the pharmaceuticals industry, Red Herring looks at how these issues are unfolding. (For more on biotech, genomics, and bioprospecting, see "Transforming big pharma.")

ADOPTED AGENTS Health officials have long been concerned about the economic problems that discourage drug research for orphan diseases. The Orphan Drug Act of 1983 gave U.S. drug companies tax incentives, grants, and seven years of market exclusivity for approved orphan products. Since then, 218 orphan drugs have been approved; another 849 are in development.

Health policy experts and scientists hope that the HGP's results will lead to new therapies for orphan diseases. But it will take several years for these technological breakthroughs to materialize, and resolving issues like patient privacy, ethics, and the patenting of genomic information may require additional time.

"Gene patenting is likely to be the big problem here," says Abbey Meyers, president of NORD, explaining that companies can hold back genetic information, demanding high prices many small biotech startups cannot afford to pay. "It is going to delay research; there is no doubt about it."

Another time-eater is the long (and costly) U.S. Food and Drug Administration approval process (see "Explained: Coping with drugs"). Ms. Meyers reports that many promising orphan drugs are shelved in the latter part of clinical trials because a small number of participants have developed adverse side effects.

Pharmacogenomics may solve the side effect problem. Pharmacogenomics uses the minute variations in DNA from one person to another to determine which patients might benefit from a drug and which are most likely to suffer its harmful side effects. "It's getting the right drug for the right patient," says George Poste, a veteran scientist who is chairman of DiaDexus (proposed Nasdaq: DDXS), a firm that translates genomic sequence data into diagnostic tests, drug targets, and novel therapies.

Among the other companies in the field are Celera Genomics (NYSE: CRA) of Rockville, Maryland, whose work mapping the human genome gives it a head start, and CuraGen (Nasdaq: CRGN) of New Haven, Connecticut, which locates genetic markers, translates the code, and finds out the activity of a gene sequence.

Bayer (OTC: BAYZY) is another competitor. It's building a library of genomic data, mostly by forging partnerships with startups and smaller pharmaceutical companies like CuraGen. To date, Bayer claims to have amassed as many as 100,000 compounds derived from genomic data. That puts it right in line with competitors, which have similar internal genomic efforts.

What will it cost to develop pharmacogenomics data? Some estimates put a price of $50,000 on identifying a single nucleotide polymorphism in one individual. Prices are likely to fall, but that will take time. At this point, it typically costs a drug company about $600 million to $700 million to develop, test, and bring to market a single drug. Pharmacogenomic data could hasten clinical drug trials, allowing researchers to design and conduct safer, more targeted trials on a particular drug. And the results of such a trial would be far more conclusive and focused than those of trials that don't use pharmacogenomic data. But sufferers of orphan diseases might not see the results for another five to ten years.

OK. Now stop thinking of it. It won't happen anytime soon, say companies and analysts in the handheld space.

'This thing is not a desktop replacement,' explains Dennis Gaughan, research director for wireless technology at AMR Research. 'It has limited memory and you will never get the user experience of a PC.'

Mr. Gaughan points out that the requisites of portability -- small screen, slim power supply -- deny users the richness and connectivity they've grown accustomed to on a desktop.  Mr. Gaughan says. 'I see handhelds not as a replacement but as more of a complement to the desktop.'

• Street Talk: No upgrades required.
• Profile: The Web cubed.
• Software management: Managing desktop mayhem.

The number of handheld users is growing. Industry leader Palm (Nasdaq : PALM) has more than 65 percent of the market, according to research firm PC Data: 10.9 million units shipped to date. Palm grabbed its lead in the market by offering a small, simple organizational device, the Palm Pilot. Competitors, including Handspring (Nasdaq : HAND), Windows Pocket PC, Symbian, Nokia (NYSE : NOK), and Ericsson (Nasdaq : ERICY), quickly followed.

Competing operating systems do more than simply boast different feature sets; there are two philosophies to handhelds. Both Palm and Handspring use the Palm OS; Palm continues to preach its mantra of simplicity in style and function. Microsoft (Nasdaq : MSFT)'s Pocket PC, on the other hand, uses a Windows-based OS and stakes out more ambitious ground: a handheld environment that integrates its desktop platform.

The Palm strategy is based on the assumption that the desktop will always be used for information creation, whereas handhelds will be used mainly for information management. The Palm Pilot's applications are always loaded; it has no rotating storage, which makes it quick and simple.

'Windows is trying to cram an entire PC into a handheld, which will be too complicated and not wearable,' says Michael Mace, Palm's chief competitive officer. 'What you end up with is a PC that's compromised.'

Mr. Mace says Palm plans to maintain its small footprint, with 8 MB of RAM and a simple, single-task application-operating system. It will work with vendors to expand the number of applications it can offer with its operating system and to continue its reach toward the Web.

Palm has also signed a developmental license with Sony and will explore entertainment technology, Bluetooth applications, and 16-bit color expansion technology, Mr. Mace says.

GROWING PANES

Microsoft's project manager for the mobile devices, Ed Suwanjindar, says Palm's expansion into wireless and its quest for more applications is an indication that consumers do in fact want more from their handhelds.

And that's where the Microsoft Pocket PC comes in, Mr. Suwanjindar says. It offers a consistent application environment with a Windows OS desktop. 'It enables a person to be more productive if the core things are designed in the same platform.' The Windows CE operating system allows Web browsing and seamless Windows transfer of email and Word documents.

Pocket PCs have more memory and processing power than Palms. 'We think 32 MB is the sweet spot, and we will go forward once memory prices go down,' Mr. Suwanjindar says.

The Windows CE operating system has enough memory and processing power to run richer Windows applications like Excel and Powerpoint, Mr. Suwanjindar says. But, he adds, 'we certainly don't want to replace the notebook. That's our bread and butter. We want to offer the mobile user the same functionality as their desktop.'

Microsoft also wants to leverage its market dominance in the desktop world into the handheld arena. The opportunity is there, Mr. Suwanjindar thinks, because the simple Palm is now outdated. 'In the early days they made a very good device, specific in function. But the user base evolves and knows the possibilities of what they can do. The simple approach is becoming more expendable.'

Of course, increased functionality and memory has a downside -- bigger, slower machines.

The Windows CE operating system may be more robust and poised to handle the desktop transition, should it come. But it -- and all the other handhelds in the market -- are still a long way from putting the desktop in your pocket.

It's no wonder there's so much confusion when it comes to broadband access: there's a slew of options for consumers to gain high-speed access to the Internet. Just consider, for example, a consumer walking into a Best Buy consumer electronics retail store. He or she would find DSL from MSN HighSpeed and Flashcom; cable television and high-speed cable Internet access from AT&T Broadband; satellite from DirecPC; and fixed wireless from Sprint (NYSE : FON). Choice is great, but it does have downsides, especially when it's hard to decipher the pros and cons.

And there are plenty of points to keep in mind when trying to determine whether a satellite dish is a better choice than DSL or a cable modem. As a last-mile solution, satellites bring the ability to reach areas outside the normal range of terrestrial landlines. Satellites also have a track record for proven security, but much of the technology is still considered young when it comes to delivering data.

With that mix of strengths and weaknesses, some analysts are projecting big growth for satellites over the next few years. Broadband satellites are expected to garner 30 percent of the Internet-access market by 2007, with 50 million users generating $15 billion in annual revenue by the end of the decade, according to Pioneer Consulting, a market research firm in Cambridge, Massachusetts.

Currently about 50 percent of broadband users get service through DSL, 25 percent through cable modems, and about 10 to 15 percent through fixed wireless technology, says Roland Van der Meer, a general partner at Comventures, a Palo Alto, California, venture capital firm that invests exclusively in communication technologies.

The remainder, about 10 percent, is expected to be filled by satellites, primarily due to their accessibility in the last mile. (For a definition of "last mile" and other high-speed-Internet-access terms, see "Special delivery.") Satellites offer the best solution for rural areas of the United States that are not yet equipped with cable or telephone infrastructure; it's not economical to run landlines to the majority of end users in nonmetropolitan areas.

PIE IN THE SKY

Satellites can reach areas that do not have existing infrastructure. Cable is limited by where the cable company has placed wire. DSL is limited to 12,000 to 18,000 feet from the phone company's central switching office. "Those who live beyond terrestrial access provide a strong customer base for the satellite industry," says Edward Sloot, a partner with the venture capital firm Sofinov. "We see it as a strong, dynamic industry with good science."

His company has explored backing at least three satellite startups, and the recurring challenge he encounters is finding a way for the satellite system to fit into the current terrestrial system. "The business model must come from a commercial perspective," says Mr. Sloot. "They need to look first at what a community needs and then how satellites can provide it."

Satellite companies have also fought to overcome the perception that they are a solution solely for rural markets. Most analysts consider satellite delivery a good match for urban and suburban settings, not just remote areas.

In the early days of dish TV, satellite was thought of as strictly for rural communities. Now it is prevalent in more densely populated areas. "This 'rural market only' perception is one that satellite companies must shrug off to earn a stake in the larger markets," says Sam Baumel, senior director of marketing for Hughes Network Systems's Consumer Division.

And much of the excitement surrounding satellites delivering last-mile Internet access can be attributed to the success over the past 25 years of very small aperture terminal (VSAT) satellite technology used in gas stations for verifying credit cards.

SKYROCKETING COSTS

But there are concerns about using satellites as the last-mile technology -- chiefly cost, but also technical issues like weather interfering with signal transmission. Consumers pay $600 to $1,000 to install a satellite dish and other equipment, says Mr. Van der Meer, adding that he doesn't believe the market for satellite Internet access is large enough to lower the price point to that of cable or DSL.

But those cost economics might be changing soon. A recent study by the Phillips Group, based in London, says new two-way satellite technology could boost the potential for satellite services being widely accepted. But the time lag in entering the market will prevent satellite technology from dominating, as both cable TV and DSL will already be well entrenched, with cable having captured the residential users, and DSL the business users.

Launch costs are yet another large expense that figures into the economics of satellite delivery, but companies like Skybridge are willing to shoulder these costs to develop their broadband distribution network. Skybridge, a company led by Alcatel (NYSE : ALA) that has headquarters in Bethesda, Maryland, is trying to deploy 80 satellites by 2003, at a cost of $6.1 billion. Of that, $2.9 billion is for the space segment, while $1.4 million is for satellite launch services and insurance. The remaining $1.8 billion is earmarked for ground and system integration.

Critics and competitors say satellite transmissions suffer from rain fade, snow load, and wind conditions. In severe snowstorms and heavy rain, for example, users may experience signal fade. What's more, heavy winds can also damage rooftop dishes.

COMMUNICATION BREAKDOWN?

Besides the day-to-day technical risks in operating a satellite network, there's also the likelihood that satellites could go offline completely or burn up in the atmosphere. In November 2000, a PanAmSat (Nasdaq : SPOT) Galaxy 7 backup satellite valued at $130 million permanently ceased transmission after an onboard system failed. That happened just days after the November 21 launch failure of a $60 million Quickbird 1 satellite, which would have provided images for the Pentagon's National Imagery and Mapping Agency and other customers. The launch was conducted by the Russian Space Agency.

Another downside of satellite technology is the inherent latency built into the transmission and reception of data signals. This is the same delay often heard when phone calls are patched through satellites, and this delay can slow data-transmission speeds.

"Whoever wins the last mile -- if they get it right -- wow, will they have the ability to sell multiple services," says Eric Hartley, founder and CEO of SatcomX, an Atlanta company that provides broadband IP content delivery and application-service-provider connectivity by satellite.

Mr. Hartley should know. The 30-year-old onetime Aspen River guide has two desktop computers, one that accesses the Internet over Asymmetrical DSL (ADSL), and the other over satellite connections. This arrangement also provides him with a lablike environment for determining what these technologies can actually deliver. Mr. Hartley says he uses the StarBand system, which began operations last April, with the assistance of strategic partners Gilat Satellite Networks (Nasdaq : GILTF), Microsoft (Nasdaq : MSFT), and Echostar Communications (Nasdaq : DISH). StarBand beams data continuously back and forth from Mr. Hartley's 36-by-24-inch rooftop dish to a satellite orbiting 22,300 miles in space. His other computer brings in broadband through a MindSpring ADSL connection. Keeping tabs on speed through simple AnalogX software, Mr. Hartley says the systems are comparable.

He has seen StarBand upstream speeds as high as 90 Kbps, with an average of 20 to 50 Kbps. Downstream speeds, Mr. Hartley says, have been as high as 1.1 Mbps, with an average of 400 to 500 Kbps. As for his ADSL line, it runs consistently at 1.2 Mbps in each direction.

By the end of 2000, Hughes Network Systems was expected to release fourth-generation technology that offers broadband return through satellites, Mr. Baumel says. It builds on the company's DirecPC system, which delivers downstream broadband by satellite and upstream broadband through phone lines. Other satellite companies that are aiming to deliver broadband include Hughes's Spaceway, Lockheed Martin (NYSE : LMT)'s Astrolink, Teledesic, and Cyberstar.

Make no mistake, demand for broadband will continue to grow. But in order to compete, satellite companies must continue to drive price points downward so the technology is as fast and affordable as DSL.

Write to ronald.recinto@redherring.com.

ADDITIONAL RESOURCES

A Columbia University professor's class on "last mile" technologies and a basic primer on broadband delivery.

Links to a trade organization that supports space technology.

Links to other trade magazine stories about satellites and the "last mile."

A short glossary of Internet-access terms:

Last mile -- the last portion of the local loop that brings broadband into the home or business.

Upstream -- the direction of transmission running from the user to the Internet. The upstream bandwidth determines the speed at which users can surf the Web or download files. Upstream speeds are usually slower with satellite technology.

Downstream -- the direction of transmission running from the Internet to the user. Downstream satellite speeds are comparable to DSL.

Geostationary -- refers to a geosynchronous satellite angle with zero inclination, so the satellite appears to hover over one spot on the earth's equator 22,237 miles above the surface.

Central office -- the location of the telephone company switching center for a given area. The central office connects the local loops to the customer premises and to the global telephone network. Often abbreviated CO.

Source: Techdictionary.com and Red Herring.

Andover, Massachusetts-based NewRiver aims to cut down on all the paper. The company's technology allows brokerage firms to automate trade confirmations and deliver documents online. The savings to financial institutions could total $7 billion, NewRiver says.

But all may not go smoothly for this upstart. Though it has veteran leadership, it faces a market that has been slow to adopt personal online investing. It also must deal with competition from entrenched firms and the challenge of enhancing Internet-focused customer service.

TECHNOLOGICAL UNDERPINNINGS

NewRiver's Prospectus Express follows an application service provider (ASP) model. Its extranet instant-retrieval system lets investors get investment information upon filing with the SEC rather than wait days or weeks for pamphlets to arrive by mail. NewRiver's technology integrates information between its servers and its clients' servers. The transition, the company claims, is transparent to users.

Since it launched Prospectus Express in April 1999, NewRiver has distributed to investors more than 2 million electronic prospectuses, annual and semiannual reports, statements of additional information, and supplements. The company charges a fee to business clients for each financial transaction. NewRiver also recently patented the process by which it maps, tags, and delivers daily EDGAR filings. EDGAR is a public database, but NewRiver's technology provides a more user-friendly interface and easier-to-read documents.

NewRiver's competition comes from several companies, including Strategic Insite, eDocs, Tumbleweed Communications (Nasdaq : TMWD), WebCenter Brigade, ClientLogic, BISYS (Nasdaq : BSYS), and Fiserv (Nasdaq : FISV). Mutual fund companies or brokerages can also provide the service to their customers. The non-ASP model requires brokerages and mutual fund companies to take on the task of updating, managing, and delivering a large amount of documentation.

Before founding NewRiver in 1995, CEO Larry Renfro, 46, worked at several financial services companies, including Allmerica Financial, State Street Bank and Trust, and Fidelity Investments. He has stocked his company with veterans he's met through the years.

NewRiver is not struggling to keep its head above water, Mr. Renfro says. The company closed $15 million in its mezzanine round of financing in November. Its initial round of $15 million in 1999 was led by a $6 million investment by Pequot Private Equity Funds. NewRiver's major investors include Lazard Frères, Taib Bank, Pequot Capital, Mariposa Investment Holdings, and Corning Technology Ventures. NewRiver landed ETrade (Nasdaq : EGRP) as a client, its biggest so far, in February 2000. In its fourth-quarter earnings report, issued October 19, 2000, ETrade said its deal with NewRiver should yield approximately $7 million in annual savings. Other NewRiver clients include Merrill Lynch, American Express, Salomon Smith Barney, The Hartford, Aetna, and Prudential Insurance.

Mr. Renfro says NewRiver will focus on turning a profit in 2001. He adds that the company is currently near break-even, and that its sound financials mean it will consider an IPO only when the time is right. "We don't have to go public to raise cash," he says.

THE FINE PRINT

Strong leadership and solid financials may not be enough to conquer a fickle market, though. Jason Briggs, the Yankee Group's senior analyst for billing and payment application strategies, says adoption of online financial services has not gone the way many people thought it would. Only about 3 to 4 percent of the U.S. population pay their bills and handle their finances online, far from the 10 to 15 percent many institutions had expected. "It's consumers not wanting to let go of the paper," Mr. Briggs says. "Most of them want to see a paper bill and file it away in a three-ring binder." And personal online investing, while increasing, also lags behind predictions. Because businesses have been moving more quickly to provide online customer service, NewRiver has broadened its model to help businesses shift their services onto the Web.

The company is also adding more account-management services. IDC analyst Aaron McPherson says that step is critical to retain and gain customers. To this end, NewRiver recently opened a call center to offer online customers real-time advice and Web support. The company is also working on an instant-messaging service by which customers can get advice while online.

NewRiver expects a broker-dealer license by early 2001. It is also anxiously awaiting e-signature legislation. If both are approved, the company will be able to offer customers complete service, from account-opening to management to electronic document delivery. Other companies are also going that direction, though. EDocs, an established account-management company, plans to introduce electronic document delivery by the first quarter of 2001. "We already deal with highly personal, perishable content," says eDocs executive vice president and founder Jim Moran. "It's relatively simple to deliver an annual report electronically."

NewRiver seems to have firm grounding, with its veteran leadership and strong financials. But the company needs to continue to grow beyond its document-delivery technology, gain more of the financial services market, and build a name for itself as a contender in online financial services.

NEWRIVER AT A GLANCE

CEO: Larry RenfroLocation: Andover MAPhone: 978/623-9900URL: www.newriver.comOwnership: PrivateFounded: 1995Employees: 200Products: E-services (ASP) for online financial managementCompetitors: Strategic Insite, eDocs, Tumbleweed, docHarborProfitable? Near break-evenFinancing: $30MInvestors: DST Systems, ETrade, MF Capital, Lazard Frères, Taib Bank, Pequot Capital, Mariposa Investment Holdings, Corning Technology Ventures, Graystone Venture Direct EquityHerring Take: The company's veteran leadership must continue to expand beyond electronic content-delivery technology. They have strong financial support and are moving in the right direction to compete.

Once in place, the projected $150 million communications facility, capable of withstanding a Level 5 hurricane, will become the latest network access point (NAP) in the United States. It will join the other four Tier 1 North American NAPs in sharing the burden of growing Internet traffic, which has been doubling every 100 days.

But this NAP is no parochial communication hub. It is designed to do much more than just provide another focal point for U.S. traffic. It is slated to become the electronic gateway that links the emerging economies of Central and South America to the United States in much the same way that major seaports helped connect global trading partners 100 years ago. But it won't be sailing solo, as this new NAP will have to compete with an NAP that BellSouth (NYSE : BLS) is building in western downtown Miami.

Behind the multimillion-dollar project is Terremark Worldwide (AMEX: TWW), an unknown that recently transformed itself from a real estate company into a telecom player.

It is footing the $80 million to $150 million bill for the technology facility. "This will help make Miami a technological center and the technology capital of Latin America," says Manuel Medina, CEO and chairman of Terremark.

LATIN LOVER

The emerging telecom market in Latin America holds great potential. A recent report by the Boston Consulting Group notes that Brazil has the largest online market in Latin America. Brazil accounts for $300 million in revenue and slightly more than half the entire regional market (for more on Brazil, see "Brazil Unplugged"). The report notes that the Mexican and Argentine markets are growing steadily and expect to generate $91 million and $82 million in sales, respectively. The analysis cautions that while sales are likely to rise 400 percent, online retailers have been struggling to ensure their infrastructure can support the demand.

Bandwidth to connect the economies will be crucial. Currently, Internet bandwidth between the Latin America/Caribbean regions and North America is estimated to be 949 Mbps, according to a report by TeleGeography. That's a small trickle compared with the bandwidth of 13,258 Mbps established between North America and Europe.

Terremark isn't trying to do everything on its own. It has joined with Telcordia Technologies, a company that designs and builds similar large data-communication centers. The stakes for Telcordia are also high: it plans to take the success of the Miami telecom facility and promote it in South America, as well as in other markets and cities. The first target is Mexico, followed by Brazil and China.

CENTER OF ATTENTION

From the outside, the six-story technology center looks like any other modern office building. But inside, state-of-the-art security devices will protect the communications fortress. Retinal scanners will control employee access. Moreover, the communications nerve center is being built on the second floor to protect the hardware from hurricane-produced floods or a terrorist attack. Redundancies in power, cooling systems, and security have been built in. All these precautions are aimed at guarding the millions of dollars in high-speed optical switches, routers, and fiber that help shuttle megabits of data to cities and homes thousands of miles away.

The first floor of the building will contain retail shops, an educational center, and parking. Global Crossing (NYSE : GX), the anchor tenant, will take the 125,000-square-foot sixth floor and use it as co-location space, says Omar Altaji, Global Crossing's vice president of sales for Latin America and the Caribbean.

The Terremark NAP also breaks ground in another sense. It's the first to be built from the ground up. The other four NAPs were squeezed into existing buildings that were retrofitted to accommodate the technology. The Terremark building has eight entry points, each with 48 conduits, to handle the amount of fiber-optic cable coming in.

To better place Miami's new NAP in the world of Internet traffic, it's useful to envision data traffic as airline routes across the United States. Currently, Internet traffic is routed through the four major NAPs: New York; Washington, D.C.; Chicago; and San Francisco. Packets of data like email pass through these major NAPs every day. This new fifth NAP is equivalent to adding a new airline hub in Miami. With the existing NAPs, if someone in Orlando, Florida, wanted to send email to Dallas, the path might be from Orlando to the NAP in Washington, D.C., and then to Dallas; with the fifth NAP in Miami, data could be routed from Orlando through Miami to Dallas. That also helps lower the cost for carriers that establish "local peering" arrangements, in which carriers exchange information along shared hardware connections.

What about international traffic patterns? Similar benefits would result. For data being sent from Lima, Peru, to Mexico City, for example, the packets would be routed through the Miami NAP instead of Washington or San Francisco, as is currently the case.

PUT UP YOUR DUKES

BellSouth, Terremark's competitor, is expected to have its project in operation by the end of 2000. BellSouth was part of the original consortium that helped land the NAP in South Florida instead of Texas. But BellSouth opted out of the then eight-member consortium and announced plans to build its own NAP in Miami. The consortium included the carriers AT&T (NYSE : T), Global Crossing, and NetRail.

BellSouth has designed its NAP to emulate a wide area network architecture, with four "nodes," or communication facilities, situated in west Miami, downtown Miami, Broward County, and Palm Beach County.

BellSouth says this more traditional approach provides a more stable and reliable scheme that won't go down in the event of a natural disaster. BellSouth differs from Terremark in that BellSouth will focus on upgrading existing facilities with stronger technology.

BellSouth claims it's been a major player in telecommunications for years, and that Terremark is a Johnny-come-lately from the world of real estate. "They are a real estate company, and their NAP is about real estate," says Stephanie Sharpe, business development manager at BellSouth. "Ours is an infrastructure NAP."

For the next several months, Terremark has leased space a few blocks from its NAP of the Americas site and is using it for an interim NAP that is expected to be up and running by the end of 2000.

The goal of the construction crews is to complete one floor each month. It's clear that the technology center will change the landscape of the area along Ninth Street. The world waits and watches to see how the new NAP will also change the landscape of the worldwide telecom scene.