Bells May Yet Be Able to Lead in Internet Market

Last month, four of the regional Bell operating companies--Bell Atlantic, Nynex, Pacific Telesis and US West--filed reports with the Federal Communications Commission pleading for regulatory relief because the growth of Internet traffic is straining their networks.

Just a few weeks later, cable giants Tele-Communications and Time Warner announced that their cable modem services were offering high-speed Internet service to select areas. The contrast between these two events--the Bell companies shrinking from opportunity and the cable companies seizing it--raises an obvious question: What went wrong?

Given the skyrocketing importance of communications and networking, it would seem natural for the Bells to reap the benefits. But despite the regional Bells’ rich experience and enviable infrastructure, they have been leapfrogged by the notoriously incompetent cable industry.

The reasons for this unlikely situation are part technical, part cultural and part political.

Examining them in detail reveals a few sobering lessons about missed opportunities and corporate lethargy. But what makes the examination more than an exercise in assigning blame is the way it exposes some hidden strengths. Weaknesses that have, up to now, kept the Baby Bells from profiting from the rapid growth of the Internet may turn out to be advantages for the same companies once the Internet matures.

Of course, if you ask someone at a Bell company why it has been so slow to jump on the Internet bandwagon, you will hear a very different story. It’s most likely that the bulk of the blame will be placed on the regulatory red tape imposed on Bells by the FCC and public utilities commissions.

It’s true that these local phone companies face a bewildering array of tariffs, strictures and limitations on what they can do and charge. But they have convincingly demonstrated an ability to get their way when they really want something--like higher rates. To understand the real reason why regulation has retarded these companies, you need to look further back.

Even before AT&T; was broken up by the Justice Department in 1984, the phone company was controlled by certain restrictions. One of the most fundamental was that AT&T; couldn’t offer “enhanced services.” It could provide the communications network but not services that manipulated the data that flowed over the network. That’s why when the creators of the Internet went to AT&T; in the late ‘60s for help in putting together their network, AT&T; turned them down cold. It wasn’t corporate arrogance, as the story normally is told, but AT&T;’s fear that the packet switching technology used by the Internet would violate the stricture against enhanced services.

Packet switching, the simple but powerful idea that makes data networks possible, divides data into short “chunks,” or packets, stamped with the address of their destination. At each step on a packet’s path, a special computer called a router looks at the packet’s destination and then sends it along on the appropriate route. This allows packets from many different computers to be interleaved, sharing a single line the way cars share a freeway lane.

The phone network, on the other hand, uses a technique called circuit switching, which resembles reserving an entire freeway lane for each conversation. This works well for voice because there are many lanes (voice doesn’t require much bandwidth) and because most conversations don’t last long. But try to use this scheme for high-speed data traffic and you quickly run into disaster.

The Bell companies are now faced with a choice: either replace their traditional circuit-switched network with something that can handle both voice and data equally well, or build a separate network just for data. Either way, they are looking at a daunting investment of time and money--at least equal to what the cable companies have had to make to transform their networks from one-way transmission to two-way.

But the Bell companies’ experience in circuit switching confers on them something the cable companies and today’s Internet service providers have no feeling for: a tradition of rigorous quality control and precise traffic engineering. After all, running a large circuit-switched network requires a lot more attention to predicted traffic levels than does running a packet-switched one. The company needs to make sure there are enough “lanes” available for peak load while still keeping costs down. The Bell companies have developed rigorous mathematical tools to do so and are now trying to apply them to data networks.

Currently, that kind of planning is a waste of time. Internet users aren’t sophisticated enough to distinguish between providers based on service qualities. But this will change. And when it does--when users grow to demand the same reliability that the phone system has traditionally offered--Bell companies may finally come back from behind.