FEATURES & COMMENTARY
New York, N.Y. — For a glimpse of the telecommunications industry’s probable future, Seth Schiesel reports that you should visit the 22nd floor of the cavernous networking center at 60 Hudson Street in Manhattan.
In one corner is a rack of equipment from Lucent Technologies transmitting 10 billion bits of information a second over optical fiber in one link and 2.5 billion in another for Deutsche Telekom. That’s impressive. But just a few feet away is a small purple box made by Cyras Systems, a California start-up, transmitting 40 billion bits a second for the German company.
Then go uptown to Rockefeller Center, to NBC headquarters. There, on the ninth floor, are perhaps millions of dollars’ worth of Internet routers from Cisco Systems. And next to them, sitting no more than 18 inches high, is a cobbled-together system using gear from Lucent and others that cost no more than $100,000 and is transmitting more than 750 million bits a second of television to Nippon Television across the street and Court TV across town.
These striking examples of next-generation technology are brought to you by an unlikely sponsor, a small private communications carrier run by a former executive of Qwest Communications International and based in Hackensack, N.J. It is called Enkido Inc.
Who? That is just the reaction of many people who work in communications. But without fanfare or high-profile venture capital, Enkido, which makes no equipment itself but is audacious in the technology it uses, may provide the best look at what communications networks will look like in five years: cheaper, simpler, more powerful.
Of course, Enkido itself may not survive that long. In its embrace of the networking world’s sharpest cutting edge, it could easily begin to bleed. But the place that Enkido (pronounced en-KEY-doh) and its chairman, Nayel Shafei, are trying to reach today is just about where most communications companies expect to end up.
Hans Roeterink, chief technology officer for Deutsche Telekom’s North American operation, said Enkido had recently provided his company “an OC-768 connection” in New York. “The connection was completed ahead of schedule and has been successfully in operation,” he said.
OC-768 is telecommunications shorthand for 40 billion bits a second. At a time when the quickest traffic flows supported by the most advanced civilian networks are 10 billion bits a second, it means Enkido may be running the fastest commercial network connection anywhere.
In the entire high-technology world, there is probably no more fevered sector than optical communications. Each week, it seems, brings the story of a small private optical equipment company, unknown outside the industry, being acquired for billions of dollars by Lucent, Cisco or Nortel Networks of Canada. Then there are the public companies. Earlier this month, in what appeared to be the biggest technology deal yet, JDS Uniphase agreed to spend $41 billion to acquire the hitherto-obscure SDL Inc., which makes lasers and other components for optical systems.
The reason the optical arena is so hot is simple. The growth of the Internet and other data communications has placed unparalleled demands on the communications infrastructure. Not only must rafts of additional capacity be built into networks, but those networks must be reconfigured to carry communications traffic in different ways. A network structure that works well for voice telephone traffic may prove horribly inefficient at carrying data.
So carriers ranging from grand dames like AT&T to debutantes like Level 3 are spending billions on advanced optical gear and other equipment often meant to make the best use of their networks for data traffic.
But perhaps no company has been as aggressive in deploying next-generation technology and embracing futuristic network designs as Enkido. In fairness, Enkido has the luxury of serving only a handful of customers, allowing it to use less proven systems. Bigger carriers have to worry more about reliability and by necessity are more conservative in their network design.
But companies like Deutsche Telekom and NBC have turned to Enkido for parts of their business because Enkido has delivered big communications pipelines quickly and cheaply. The Naval Research Laboratory in Washington, an agency of the Defense Department, is working with Enkido as it tries to jump-start the next generation of optical networks in connection with the Defense Advanced Research Projects Agency, or Darpa.
“We’re not trying to pick winners; we’re trying to deal with anyone who can provide some sort of capability,” said Hank Dardy, the laboratory’s chief scientist. “I don’t know where Nayel’s going totally, but he’s helped to coordinate several things for us in terms of getting us the right components and connections for the bandwidth we need.”
At a certain level of abstraction, it’s fairly simple to describe where Mr. Shafei wants to go.
“Big fat pipes for big fat customers, that’s our business,” said Mr. Shafei, who until the spring of 1999 was an executive vice president at Qwest. “When I left Qwest, I said, ‘I can put together an end-to-end network, both long-haul and local loop, that will be competing only with Fed-Ex, taking packets door-to-door.’ ”
That sounds awfully similar to a lot of companies out there, companies that have far more extensive networks. But it is the way Enkido is pursuing its vision that makes it interesting.
First, a bit of background. Most communications networks today are a jumble of languages and systems. These different protocols and technologies are generally layered upon one another, or within one another. Imagine a shipment of vases from London to New York. The vases may be packed into crates, the crates placed into a shipping container and the container put on a ship. The vases are all the customers in New York care about, but the crates, container and ship are all required to get them there safely.
Now think of optical fiber cables and the equipment that lights them as the ocean. In most traditional communications networks, the role of the ship is played by a technology known as Sonet, for synchronous optical network.
Sonet is reliable, but it can also be inefficient, not only because a significant amount of communications capacity must be set aside for “protection” in case of a failure, but also because at the most basic level Sonet packages information in envelopes of 52 million bits of information a second.
On top of Sonet, many carriers deploy a technology known as A.T.M., for asynchronous transfer mode. Think of A.T.M. as the big containers. Containerizing was a revolution for shipping because containers’ uniform size allows them to be shipped easily around the world, by boat or train, without regard for what is inside. A.T.M. performs a similar function by packaging voice, video or data communications into uniform “cells” of 53 bytes each.
And then, within the shipping containers, within the A.T.M. cells, are the actual services that people use: telephone calls, Internet services, or other data services. There are obviously many other ways to configure a network, but many if not most of North America’s big carriers use or are moving to this structure of services.
Put simply, what Enkido wants to do is eliminate the A.T.M. and Sonet layers, which seems perhaps like the equivalent of throwing those crates of vases into the Thames and expecting them to wash up in Manhattan in two days. If it works, huge costs and complications have been avoided.
It sounds crazy, but that is just where many communications experts expect the networking world to end up, with some of the many communications languages becoming extinct or being radically modified and Internet protocol eventually riding directly over wavelengths of light.
“If I were starting now I would go straight to I.P.,” said Pushpendra Mohta, a former AT&T vice president who is now an entrepreneur-in-residence at Benchmark Capital, the prominent venture capital firm in Menlo Park, Calif., and who is just starting a new wireless data company. “If I were starting something now I would make that bet, to get up straight onto wavelengths.”
Muayyad Al-Chalabi, director for core communications switching analysis at RHK, a consulting firm in South San Francisco, Calif., said: “The direction is higher and higher speed and elimination of layers. There’s no question about that. The question is who can support millions and millions of users.”
And it is there, in the question of what communications executives call “scalability,” that Enkido may face its biggest challenge. It is easy running what some might call a science project of a network when you have only a handful of customers. But when the number of customers takes on a fair number of zeroes, the technology must often be robust enough to take care of itself.
For now, Enkido is aiming at the fattest of customers: financial services firms, entertainment companies, other carriers and the government.
After leaving Qwest last year, Mr. Shafei put a few million of his own dollars into starting the company and then began approaching prospective customers and companies that could sell him raw optical cables across the nation and within the New York area.
With the money he got up front from customers, led by NBC, he was able to begin buying space for Enkido on optical cables owned by a panoply of local utilities and by companies running cables across the continent.
This year, Mr. Shafei expects Enkido to generate more than $50 million in revenue. As advisers he has signed up Raj Reddy, a professor of computer science at Carnegie Mellon University and co-chairman of the President’s Information Technology Advisory Committee, and Robert E. Kahn, who spent 13 years at Darpa before becoming president of the Corporation for National Research Initiatives, a nonprofit research group in Reston, Va.
More than 90 percent of Enkido’s equity, however, remains with Mr. Shafei. And he does not intend to share any of it with venture capitalists, whom he views derisively.
“Why on earth do I need to go have a monkey sitting on my shoulder, telling me ‘Do this, do that,’ and they don’t know anything about the industry,” Mr. Shafei said. “I’m not a miser with equity, but I don’t like monkeys.”
Instead, Mr. Shafei says he has arranged a line of credit from commercial banks worth $100 million to $200 million and added that he would consider offering high-yield debt financing, or junk bonds, in the future.
In his financial plans, as in his technical ones, Mr. Shafei is taking the path untrodden.
“You have dinosaurs out there in this business, and then you have dinosaurs with space helmets on, but they’re still dinosaurs,” he said. “We’re a whole new breed in the evolution ladder. We have five fingers and five toes.”