By Jeffrey Krauss, surfing the telephone net and President of Telecommunications and Technology Policy

Now the telephone industry has raised a new complaint—Internet subscribers are screwing up telephone networks because the networks weren't designed for Internet connections. Phone companies just can't get it through their monopoly mindset that when customer demand patterns change, the service has to change. They don't understand that in the new world of competition, the ability to offer the service that customers want will be the difference between success and failure.

The access charge issue

An ISP is something like a long distance phone company, in that it gives subscribers access to a long distance communications network. Both the ISP and the long distance carrier incorporate local phone connections in their overall network. Both pay the local phone company for, in effect, reselling the capacity of the local phone network. But they pay different rates. The ISP pays the same rate as any business that uses the phone network to make local calls. The long distance carrier pays "access charges." Access charges are much higher than local phone rates, as high as five cents per minute.

Access charges originated in the early 1980s as an FCC-sanctioned way for long distance service to subsidize local phone service. In 1987, well before the Internet became popular, the phone companies wanted the FCC to apply access charges to enhanced service packet data networks like Telenet and Tymnet, but Congress objected, and everyone lost interest in the issue. Now, with the rise of Internet telephony, the phone companies are at it again. They want their subsidies.

The current expectation is that the FCC will overhaul access charges in the 1997–'98 time frame. ISPs will have to pay access charges. But the rates will be knocked down from five cents per minute to a few tenths of a cent.

The network design issue

The new argument from the phone companies is that Internet traffic is different from voice traffic, so different that it could crash the phone network.

Wait a minute. Let's take a closer look at this argument. Sure, phone call durations might be longer when I surf the 'Net than when I talk to my mother in Florida, but what are the cost and engineering implications?

By the way, even the phone industry agrees that voice call durations have increased over the years. It used to be that an average phone call would last four minutes; now, the phone industry uses nine minutes as the average holding time.

My local loop and the connection to the local phone switch are dedicated to my phone number, whether I use it or not. The cost is independent of the duration of my calls.

The local switch gives me dialtone, interprets my dial pulses and arranges for a path through the network to be established. But that happens only when the call is dialed. After that, the switch circuitry goes on to set up the next call. Most of the usage of the telephone switch circuitry is independent of call duration.

But the phone industry is right about one element. There is a part of the network that is affected by call duration, the trunking between local switches. This is because the phone network assigns a circuit full time to a connection between switches, even though you might be sending and receiving data packets only sporadically.

But rather than figuring out how these trunks can be shared among multiple Internet connections so they can be used more efficiently, the phone companies are now spreading scare stories about the Internet causing a possible "meltdown" of the phone network.

The monopoly mindset

It's the same old story, but now with the local phone monopolies instead of a single nationwide Ma Bell. "Here's our network and our service; you must tailor your demand so it fits what we have."

Let's look at another part of the Internet, the intercity high-speed data circuits between ISPs. One of the leading suppliers of this service is MCI. MCI took its intercity fiber network, which was installed for voice telephone traffic, and figured out how to use it efficiently for high-speed data.

MCI saw the Internet as an opportunity, not a threat, and modified its network design to satisfy user needs. And the company earned a profit while doing it.

The phone company engineering philosophy has been stability, not agility. But the world is changing. Stability fits quite nicely with a monopoly environment. But "brand loyalty" is on its way out. Have you noticed that in the cable industry? It applies to the phone industry as well.

In the world of competition, companies must change, or die.