DESPITE ALL THE TALK of competition in the high-speed access space since the passage of the Telecommunications Act of 1996, there haven't been many competitive offerings available to most enterprises, at least not in terms of the last-mile connection to the nearest Internet point of presence. For most businesses, the only real choice has been the local telephone company's venerable T1 connection -- a couple dozen pairs of ordinary copper phone lines aggregated into a bundle that offers speeds of no more than 1.5Mbps. The only options generally have been a fractional T1 or a multiple T1 at correspondingly lower and higher monthly tariffs -- in other words, not much in the way of alternatives.
Seeking to fill the need for options, several challengers have appeared recently. DSL was supposed to offer T1 speeds at a much lower price, but real-world implementations seldom matched T1 performance. And when they did deliver the speed, their prices were apt to be equally elevated. In any event, DSL services as a whole have yet to demonstrate enterprise-class reliability and, moreover, the services have been plagued with well-publicized security lapses as well as disquieting business failures among the providers. Other high-speed access technologies -- such as broadband wireless, cable data, two-way satellite and fiber to the building -- either haven't happened yet or at least haven't been available in the bulk of business districts.
Dawn of the D-LECs
Then, with little advance publicity, a host of new companies called variously data-centric local exchanges (D-LECS) and building local exchanges (B-LECs), derived from the fact that many of these companies bring their services to multiple companies in a building, began to challenge the incumbents with services that offered throughputs in the tens or even hundreds of megabits per second at prices in the hundreds or low thousands of dollars per month. The list of insurgents is long: Dallas-based Allied Riser Communications (ARC), Cogent Communications in Washington, D.C., OnSite Access in New York City, Terabeam in Seattle, New York City-based Winstar and San Francisco-based Yipes, just to name a few.
While none of these upstarts share identical service offerings or business plans, all place their emphasis on high-speed data as opposed to voice -- hence the name, data-centric local exchanges. But that's far from all they do. Quite a number, including OnSite Access, Winstar and Yipes Communications, offer voice as well -- generally in a bundle combining local and long-distance calls with high-speed Internet access and LAN extension. Some of the companies provide Web and application hosting. A number, including ARC, even sell LAN consulting services.
Diverging service offerings aside, all of the companies resemble one another not only in their data-centric points of view, but in providing impressive speeds at cutthroat prices. Plus they each seek to offer very fast time to service, thus attacking the two biggest perceived weaknesses of the incumbent providers -- high cost and long delays for implementation.
Of these two key attributes, the cost-performance ratio is probabbly the more attractive to most subscribers. Santa Clara, Calif.-based voice-over-IP system provider TollBridge Technologies, another Silicon Valley early adopter of the Yipes service, reports a similar experience. "We couldn't believe the price," says Agnes Imregh, former vice president of marketing. "We jumped at it, and so far it's been great."
Speedy arrival times have also won over a number of the initial subscribers to these new services. Denis Kavanagh, chief operations officer for David Lawrence & Associates, a New York City-based executive search company, was impressed: "We needed a second T1 line, and AT&T told us [it would take] another 60 days after they'd already strung us along for several weeks. OnSite Access said four weeks, and they met their deadline."
Matthew Kesner, CIO for the law firm Fenwick & West in Palo Alto, Calif., was similarly happy with the speed of provisioning displayed by Yipes. "They offered to get us any increment of 10 megabits within 10 days of our placing the order. And they delivered," he says.
Such enthusiastic customers aren't alone, and practically all of the larger D-LECs and B-LECs have been accruing new accounts and territories at a dizzying pace (though they still make up only a tiny fraction of the overall market for high-speed business access). This leads to a logical question: How can they price their broadband services so far below those of the incumbents? And how are they likely to fare once the incumbents become aroused and when multiple D-LECs begin to compete in the same market spaces?
Guerrilla Access Strategies
"Most of the new companies in this space pursue broadly similar strategies," observes Kevin Mitchell, directing analyst for service provider networks at San Jose-based Infonetics Research and an authority on the D-LEC/B-LEC phenomenon. "They're leasing dark fiber [fiber-optic cable that doesn't have communications equipment attached to it] -- in many cases, from Metromedia Fiber Network [in White Plains, N.Y.] -- and they're setting up the networks as large gigabit Ethernet LANs." The gigabit Ethernet equipment is far less expensive than the infrastructure for a synchronous optical network (SONET) or asynchronous transfer mode (ATM) networks, giving these companies a way to undersell the incumbents.
"Personally, I think the business plan is very sound because it is in harmony with long-term trends in network evolution," Mitchell says. His view of network evolution encompasses simplified design and administration, easier implementation and more flexible service offerings.
Gigabit Ethernet at the core and pure IP traffic over the network also simplify network connections for the enterprise user, because they permit a direct termination at an Internet port at the enterprise's own switch or router. And that certainly works for Mark Dumic, manager of network systems and telecommunications at Swarthmore College in Swarthmore, Pa., and an enthusiastic user of the Yipes service: "The pure Ethernet transport is ideal with us because we're linked up with the networks at Bryn Mawr and Haverford, which have fast Ethernets as well."
On the other hand, the questions surrounding a gigabit Ethernet architecture may make some IT managers hesitate. "You don't have the proven subsecond restoration capabilities of SONET or real carrier class equipment," notes Mitchell. And both AirFiber, a San Diego-based manufacturer of free-air laser infrastructure equipment, and North Andover, Mass.-based Quantum Bridge, which makes low-priced passive optical systems, espouse ATM.
Beam It Up
This all assumes, of course, that the network is Ethernet-based. In the relatively small number of D-LECs using wireless access technologies, the carrier standard ATM is just as likely to be used. Winstar uses ATM-based microwave equipment, while Tellaire, a Dallas-based startup using free-air laser beams to connect its nnetwork, also uses an ATM core. AirFiber, a San Diego-based manufacturer of free-air laser infrastructure equipment also espouses ATM.
In each of these cases the higher cost of ATM switching equipment is purportedly offset by the lower cost of the physical medium for the network -- air. But once again, reliability questions inevitably arise.
While some wireless providers, including Winstar, claim five-nines reliability, analysts in the field note that current generation local multipoint distribution services equipment in the field hasn't been around long enough to make such assertions entirely convincing. And the free-air laser systems are apt to inspire even more grave doubts because the equipment has gained a reputation, openly admitted by most manufacturers, for link failure in the presence of heavy fog. However, at least a couple of the newer manufacturers including Richmond, Canada-based Sona Communications and Terabeam, which, incidentally, is also a carrier, are touting new laser technology that supposedly reduces the weather dependency of the systems.
Ultimately, however, the biggest factor determining the fate of the D-LECs may not be the strength of their respective technologies but the reactions of the well-entrenched and well-financed incumbents. At the moment, most of these competitors are simply watching and waiting, though several long-distance capacity suppliers such as Level 3 Communications, in Broomfield, Colo., and Denver-based Qwest Communications are reportedly contemplating D-LEC-like services of their own. "The RBOCs [regional Bell operating companies] will wait this out for a while, and when it's taken enough of their T1 business, they'll get into it themselves just like they did with DSL once it proved to be a real market. And once the pioneers start failing, companies like Yipes may not be sustainable in the long run," says Solomon Wong, executive vice president of Ottawa-based Akara, a manufacturer of optical equipment for the metro space and supplier to both the incumbents and D-LECs. "We'll just have to see."
This story, "High Speed, Low Cost" was originally published by CIO.