PROFILE; A flash of light.
Telephony, March 20, 2000 pNA
By Biagi, Susan
New services and an increasing reliance on communications are pushing the need for high bandwidth to the edge of the network. MFN is already there
After a quarter of a century installing electrical cabling in New York, Stephen Garofalo saw the light. Or at least the need for it. "I had a communications company and a cable installation company for 25 years. I got to see it from both ends: the construction and the services. And there was a blending of the two industries," he recalls.
His observations on future technology needs and service growth solidified his vision for a new company in 1992. "Communications was concerned with the invention of fiber optic cables - it's lasers now - that allowed you to transport 2000 times more communications on the same pair of cables the size of a hair on your head," Garofalo says. "The same labor was involved [in laying fiber], and you could deploy 2000 times more capacity." It was a no-brainer.
He founded Metromedia Fiber Network in April 1993, with the goal of bringing fiber to the customer. The White Plains, N.Y.-based company is accelerating its pace down the photonic path, and there is no end in sight, says Garofalo, who is CEO. "I'm taking about distributing the photonic signal - bringing a strand of fiber to every home, every office building, every floor where communications will occur. It's a project that will take 20 years to complete, at least - maybe 30 years."
Garofalo assembled an expert team of industry veterans to help MFN meets its goal. Among them are Howard Finkelstein and Nick Tanzi. Finkelstein has held a host of positions since coming on board in April 1997, and he recently was promoted to vice chairman. Tanzi left his post with Fujitsu's enterprise division to join MFN in 1997 and was named president and chief operating officer earlier this year.
The team has led MFN up a steady growth path from 1996 revenues of $236,000 to 1999 revenues (as of Sept. 30) of $10.72 million. Since MFN's IPO on Oct. 29, 1997, the stock has climbed steadily from $21 3/8 per share, has had several splits and now trades in the low $80s. A market capitalization of $17 billion and a stock appreciation of 55% in 2000 - up 186% from 1999 and up 703% from 1998 - puts it squarely in the league of the dot-coms.
MFN has the energy of a dot-com, too. Tanzi and Finkelstein support Garofalo's vision; their passion is evident in their retelling of his fiber start. As he tore up the streets of Manhattan as an electrical contractor, Garofalo realized the significance of installing fiber, which customers were asking for, Tanzi says. "Nynex laughed him out the door."
That didn't stop Garofalo, however. Since he was digging up the streets, it was cost-effective to add fiber to his build, Finkelstein says. "He put more fiber in that very expensive build to reduce the per-unit cost of the fiber itself," he says. "Then he could offer raw data infrastructure to more carriers and corporations directly."
Garofalo correctly envisioned "a bandwidth intenseness in the local market for which he could provide dark fiber as a product or service in the market. He hocked his house, maxed out his credit cards and got one of the few exclusive franchises that permitted you to lay fiber," Finkelstein says.
The first network build was successful as carriers and large data users such as financial institutions signed on for service. With the ratification of the Telecom Act of 1996, distribution of data became an issue: Once all the data was transported between cities, how would it get to the end customer? Through local loops - copper and fiber.
Garofalo's initial fiber build "was years ahead of its time," Finkelstein says.
City living
MFN's approach is to build the fiber pipe in highly populated buildings in metropolitan markets. Instead of building metro rings throughout a city, it handpicks specific buildings and links them strategically to the appropriate central offices (COs).
"We pay absolute attention to not where COs are but where the customers are," Tanzi explains.
The company plots every building within a 100-mile radius that houses a Fortune 100 customer or meets a square-footage requirement. "We use our own algorithms to estimate the bandwidth going into those facilities. Then we get the rights of way, so our rights of way sit on the doorstep of every building," Tanzi says. "Our networks - they're not nice rings. They zig and they zag."
MFN will build beyond a city, but only along populated paths. It can be a slow process. "In general, it takes us two years or more from start to finish - a little more in bigger markets," Finkelstein says. "We've got to design where we're going. It's not a railroad where we marry scores and scores of rights of way. We go through many different municipalities and franchises, and we need permits for all of that."
Unlike other companies building out metro fiber rings, MFN focused on fiber connectivity, not bandwidth usage. The strategy of this company turns metered bandwidth on its ear, Finkelstein says.
The company charges by fiber strand and distance. "Our product is providing a strand of fiber for so many miles," Finkelstein says. "We grow the business by growing the distribution." That translates to more fiber miles.
He likens MFN's approach to charging trucking companies $1000 for a gas pump and letting them pump as much gas as they want for free. Service providers and some corporate customers lease fiber from MFN but can use it however they like. Some add wave division multiplexing (WDM) equipment; others use the fiber to link LANs. MFN doesn't care. Its network operates independently of technology, service and equipment.
"Our business is end-to-end distribution. We're taking the cost of bandwidth out of the equation," Finkelstein says.
Today, MFN has operational networks in five cities: Chicago; Dallas/Fort Worth, Texas; New York; Philadelphia; and Washington. Thirteen others are under construction. Across the pond, 16 European networks are under construction, and 33 others will be announced before the project is complete. When finished, the network will include 67 markets worldwide and be comprised of 1 million fiber miles.
When MFN rips up the streets, it lays 8 to 12 conduits, populating them with 432 to 864 fibers. The fiber can support Sonet, WDM, ATM, frame relay, gigabit Ethernet and video transport at OC-3 to OC-192 rates. Its network operations center in New York monitors the links but not the services riding upon them.
"We don't provide any services to our customers," Finkelstein says. "It's all leasing infrastructure - to a carrier or very large private user of infrastructure. They develop their own services on it."
The fiber can transport up to 2 Tb/s. That might be overkill today, but it won't be for long, Tanzi says. "We tell investors to step forward three or four years when Internet access devices are on the desktop and are able to access real-time and store rich media content. [For that], you absolutely need the underlying power of fiber optics."
Data play
But just having the metro pipes isn't enough for MFN. In September 1999, the company spent $1.8 billion to acquire AboveNet Communications, an Internet connectivity provider that had purchased dark fiber from MFN. The company is in the process of integrating the AboveNet national data backbone with MFN's fiber optic network.
Dave Rand formed AboveNet as an Internet infrastructure company, similar in concept to MFN's pipe idea. He now serves as chief technology officer for the combined company.
The AboveNet network determines the least congested path for a packet to travel. Its "one-hop network" concept could be translated as one hop onto the AboveNet backbone, which transports the data as close to the end user as possible before it hops off. While the data travels across its network, AboveNet can control the latency and performance. Many network providers support the theory that the less time a packet travels on its network, the less it costs the provider. That might be true, but it results in best effort, multi-hop data services, Tanzi says.
"We carry it to the farthest point on our network, closest to where it's destined for," Tanzi says. "We change the paradigm to one of network infrastructure that is built for performance, which is critical to commerce, entertainment, education."
In fact, AT&T provides AboveNet with detailed routing information to carry its traffic. That enables AboveNet to follow deaggregated routes. "For other peers, they only advertise the aggregate," Rand says. "We get more control. We redirect the traffic in the face of congestion."
Having the data piece puts MFN in an enviable position as data transport rides the hockey-stick curve. Customers are experiencing 10% compounded annual growth rates in data traffic, Rand says.
With expected growth rates like that, he says, "the optical networking side of the business is pretty vital." But ISPs can use MFN's network to expand rapidly and to expand to places otherwise cost-prohibitive to serve. And it improves time to market, Rand says. An ISP might wait 14 months from the time of order to get a circuit, which stymies its growth or forces conjecture rather than reliable future planning.
"Getting a circuit in 14 weeks is unacceptable. Fourteen months is ridiculous," Rand says. "ISPs are not equipped to grow quickly. But now an ISP without fiber can upgrade literally in an afternoon."
By maintaining a focus on transmission instead of services, Rand believes ISPs just need to step up to the fiber plate. That's one challenge incumbent upon MFN. "It's education for the ISPs to realize that the fiber is available. No one else is making that fiber available to them," he says.
MFN's fiber links to COs and to a slew of Internet service exchanges (ISXs), which it acquired with the purchase of AboveNet. The network can support "multiservice content on a pipe in a market that is bigger than any pipe that exists," Tanzi says.
MFN announced in February a $1.4 billion expansion to link all the ISX facilities in North America and Europe and add 14 more and boost its European co-location space by 1 million square feet.
"Our view of the world is the infrastructure that exists today, namely the public network, from a business-to-business perspective becomes irrelevant in the places that we're building networks," Tanzi says.
A friend in need...
Service providers apparently agree with Tanzi's assessment. They are signing on in droves. The fiber operator boasts an A-list of customers that includes Allegiance Telecom, Winstar Communications, PSINet, America Online and CBS News. It also counts Williams Communications as a customer, noting that the metro fiber play is synergistic with Williams' long-haul network.
As of January, MFN had signed contracts totaling $2 billion - some for cities that aren't yet built - and subsequent contracts have pushed that number up even higher. Not bad for a fairly young, niche-strategy player.
"Customers follow us from market to market [because] they're comfortable with the infrastructure we're building," Tanzi says. "We're not a [competitive local exchange carrier] putting extra fiber in the ground and selling it off to defray capital costs. This is our core business."
Companies such as Allegiance and Winstar have signed on for future markets; in fact, sometimes they request that MFN enter a specific market for them. Winstar appears to be one of MFN's biggest fans. In October, the wireless provider signed a $300 million deal for dark fiber in most of MFN's U.S. and European cities. The operator uses fiber to link its switches and hubs and aggregate its fixed wireless traffic.
"We've committed to buy in advance of the build," says Bill Rouhana, CEO of Winstar. "They showed us what they are planning to do. Many of those cities we indicated a strong interest in. We'll be a big customer going into those cities. It's more cost-effective than if we did it ourselves, and it's done more quickly."
The companies have extensive mutual agreements. "We are in 50 markets with them and in 10 internationally," Rouhana says. "We buy dark fiber from them, and they buy lit capacity from us for AboveNet. We buy the heck out of them." Winstar's agreements with MFN add up to $350 million. Other strategic partnerships include a $2 billion financing relationship with Lucent Technologies and a $400 million deal with Williams.
The AboveNet relationship with Winstar existed prior to its MFN acquisition, but Rouhana was pleased with the result. "I don't mean to be a Pollyanna on this, but they have a great company. We love to do business with them," he says. "We're two very aggressive companies, and we want to do things as quickly as possible. Both of our styles are fast-paced."
Allegiance also counts on MFN for 16 of its 20 markets served. The local provider obtains fiber from Level 3 Communications in other areas, including a long-haul link between Boston and Washington. Royce Holland, chairman and CEO of Allegiance, considers MFN a preferred supplier. "We like to see them go into additional markets," he says. "We like working with them. They are very responsive to what we're looking for."
Despite Nynex's early chuckle, Bell Atlantic forged an unusual agreement with MFN, allowing the company to co-locate in its COs in exchange for fiber.
The Bell Atlantic agreement is "a reasonable quid pro quo," says Carl Garland, principal analyst network services with Current Analysis. "Metromedia has a lot of dark fiber, and not having to fight Bell Atlantic for co-location agreements is certainly a major plus." SBC Communications and Williams have a similar arrangement.
Potential conflict?
MFN's aggressive network build is no problem, nor is the technology-agnostic plan, but the company eventually may run into problems with its fiber-only offering, Garland says. "The only trouble with that approach is that it gives the advantage of pushing for a low-cost position," he says. "The downside - and the upside - for all wholesalers is if you are the cost leader, you'll always win in that market. As bandwidth moves toward being a commodity, companies like Metromedia and Williams are basically going to be trading bandwidth."
Although MFN avoids service offerings, it has teamed with Nortel Networks to offer dense WDM services to carriers. A similar arrangement with Cisco Systems provides carriers with gigabit Ethernet capabilities. The services are integrated by the provider customer, not MFN or the vendor, Finkelstein says.
But that could be a hard position to maintain. MFN already offers fiber directly to large enterprise customers, and the shift to offer services might be only a few years away. It could offer managed data services, DSL, e-commerce, virtual private networking or firewall services. If so, its large enterprise customers could bypass the middleman - an AT&T or SBC for example - and buy services directly from the fiber provider. AboveNet, Garland argues, already offers managed data and voice services to smaller ISPs.
"That would be moving up the value chain," Garland says. All fiber and bandwidth providers are in the same boat. As fiber providers build out more and let third parties provide value-added services, "their total revenues might go up, but their margins are going to get slimmer. There are richer pickings in the service sector," he says.
One alternative could be to offer a limited pool of services and avoid more complex items, such as billing and customer care. But a move like that could introduce channel conflict and scare off fiber buyers.
Because margins likely will tighten, entering the service space would be a logical move. "That is speculation," Garland says. "They would deny it until the cows come home." And they have. MFN and Williams each have gone on record refuting insinuations that they have plans to serve end users directly. Garland acquiesces, but adds, "that's a logical step down the road."
Another alternative to the margin squeeze is to merge with a provider. Williams could be a good fit for MFN; it operates a substantial long-haul network but lacks customer access. If both companies climb aboard the services train, a union might benefit both. "Metromedia plus Williams would be a poor man's version of Bell Atlantic going after GTE," Garland says. Still, he says, it's a long shot.
Allegiance's Holland dismisses the idea that MFN might need to move into the services game. "Metromedia does very well at what its doing. Going into the service business is a different move. For one thing, the number of employees you need increases exponentially.
"They are successful, customer responsive and they've built enough networks now that they're good at it. I kid with Howard Finkelstein and call them the arms merchants to all of the belligerents - the providers like us," Holland says.
It's a matter of economics, he adds. The number of fibers MFN installs and the corresponding number of calls the fiber supports allow the company to build its business on fiber sales alone. "They still have a nice margin," Holland says. "That's the reason their business model works."
On the light course
Today, no one at MFN is interested in serving end customers. They clearly state that managed services are not in their plans. For now, MFN is focusing on building its networks and planning for more. The AboveNet integration will consume its remaining energies, and Garofalo is certain that will pay off.
"Clearly the Internet today is not the Internet you'll see tomorrow. A renaissance is being created. Full-motion Internet is not far off," he says. Without the fiber infrastructure, he says, the company was putting the cart before the horse. The existing last-mile copper infrastructure simply can't handle huge amounts of data. With fiber access to the desktop, developers can create super high-bandwidth applications. In short order, customers will be able to download video clips from a meeting or press conference or download movies on demand.
But delivery hinges on reaching the end customer. "We want to create a pure photonic network with fiber to every [customer]," Garofalo says. "That's what we set out to do, and we've been doing it.
------------ |
