Online Extra: "This Isn't The Old Ma Bell"

AT&T tech honcho Hossein Eslambolchi explains why the telecom giant is no longer a pipes-and-plumbing company

The communications industry has long been known for its conservative approach to management -- in large part because its executives grew up working in a regulated industry, protected from the vagaries of open competition. But Hossein Eslambolchi, AT&T's chief technology officer, chief information officer, and president of its Global Networking Technology Services business, stands out among the "Bell Heads." Known for having more than 400 patents to his name and his penchant for immediately responding to e-mail, he bubbles over with enthusiasm regarding the future of AT&T (T ) and the Internet.

He has certainly had a major role in the recent development of both. In 1998, he pushed through an ambitious plan to forge Ma Bell's many specialized networks into a single, more efficient system based on Internet technology. That way, it could provide everything from Web-hosting to videoconferencing to Net-based calling, without duplicative spending on the underlying networking gear.

The result: AT&T's network now handles more Net traffic than any other -- some 1.8 petabytes of digital fare a day. BusinessWeek's Peter Burrows spoke with Eslambolchi three times in recent months about the state of AT&T, the Internet, and the tech sector in general. Here are edited excerpts of their conversations:

Q: AT&T has spent billions in recent years to revamp its network with Internet-style technology -- this at a time when the company's long-distance revenues were plummeting. Why was this necessary? What's the megatrend behind your thinking?


Back in the 1980s, all the intelligence [in the communications world] was in the central network. You could plug one of those plain old black phones into any phone jack and dial an 800 number, and some big phone switch out there would make the necessary connections -- without the phone knowing much of anything. Then, in the 1990s, all the pundits claimed that we would go to smart devices (such as PCs, PDAs, and cell phones), all communicating over dumb pipes. We went from one end of the pendulum to the other.

Q: What's wrong with that?


You need to have some smarts in the network. Think about security. If you're trying to connect 1 billion PCs, it becomes unmanageable [to try to make sure each one is protected all the time]. CTOs aren't getting the productivity they should be getting. So we have to build networks that can predict viruses and worms [before millions of those unprotected PCs are infected].

That's why we've been pushing a philosophy that says you can't have have all the smarts in the network or in the devices. We need smart devices, communicating over intelligent networks -- in other words, the best of both worlds. That's what our strategy is all about.

Q: How are you doing at building this new network?


Back in 1998, I told [then CEO] Michael Armstrong that we had to go to IP [Internet protocol]. I told him that IP is like Pac-Man, it would eat everything, including [the voice business]. Back then, we were number four or five in terms of the amount of IP traffic we handled. But you don't win the New York marathon when you cross the Verrazano bridge. We started late, but we'll finish first.

Q: Most readers know that AT&T has essentially pulled out of consumer long-distance, which was its main business for so many decades. [Earlier this year, AT&T said it would cut all marketing efforts to sign up new customers.] How will this network be used to remake the company?


We're turning AT&T into a software-based company, not a pipes-and-plumbing company. This isn't the old Ma Bell.

Q: Is that why you're partnering with Intel (INTC )?


We want to create a software infrastructure so that any company can plug any application on any device into the network [and the network will give the proper security and bandwidth to ensure good performance]. And who has the best, most global, most intelligent global network? Clearly it's AT&T. And Intel has the world's best computing technology. It's in almost every PC, and increasingly in plasma TVs and lots of other devices. Together, we can build industry standards for the 21st century.

Q: Is this an attempt to embed proprietary hooks to AT&T services into Intel's chips, so that AT&T's network services work better than those of other carriers?


There is nothing proprietary about this. That's one of my main passions. We're trying to create an open industry consortium, so that all of these devices will work with any of the services. At that point, you compete on the performance and reach of your network.

Q: Why are you working with Intel to push WiMax [a technology that would let computers connect wirelessly to the Net from as far as 18 miles from a WiMax tower]?


In the U.S., there are 270,000 office buildings. AT&T has direct access to only 7,000 [that have contracted to have AT&T lay fiber links directly to their premises]. We paid $8.5 billion last year to local-exchange carriers [in access charges for use of the phone lines to these businesses and homes]. WiMax could get us to the 200,000-plus buildings we don't connect to directly today. It's a wireless technology that's showing a significant amount of promise.

Q: But other carriers, from British Telecom (BT ) to NTT (NTT ) to Verizon (VZ ), are also making big investments in cutting-edge networks. Will your lead last?


They're trying to design a Space Shuttle to go to the moon. We're already on Mars, and we're trying to get to Pluto. That's the difference.

    Before it's here, it's on the Bloomberg Terminal.