Global Strategy Business

January 1998

It's The Stupid Network

To the big telephone and telecommunications companies, IP telephony is more than just another way of carrying voice traffic, or a worrying technique which threatens to divert 5% or even 10% of their voice calls away from circuit switched systems. It is a key part of a range of technologies which threaten to overthrow the dominant technological paradigm of the telephone industry. Many of the reasons were described in a recent paper from David Isenberg, a senior researcher at the Opportunity Discovery Department at AT&T Labs. The paper has been described by Harry Newton, president of Telecom Library, as "the most controversial to come out of the telephone industry ever." It is clear Isenberg's views are not those of AT&T.

According to Isenberg, telephone companies are investing heavily to make their networks more intelligent. This means they are able to offer services such as automatic re-routing, support for calling cards, giving the caller choices before the call is finished, and supporting a variety of payment options. They do this using expensive computing equipment intertwined into the network architecture. But the intelligent network is based on four assumptions: that infrastructure is expensive and needs to be shared; that voice dominates; that the main underlying technology is circuit switched; and that the telephone company is in control of its network.

None of these assumptions is currently holding true, however. Infrastructure costs have fallen a thousand times over; data traffic is overtaking voice; there are now many different alternative telecoms networks; and the Internet is pulling control of the network away from the center point to the end user. "The 'intelligent network' is being superseded by the 'stupid network', with nothing but dumb transports in the middle," says Isenberg. With a stupid network, such as the Internet which merely routes packets, the end user can request what kind of line quality is required, put together conference calls from the desktop, monitor billing and link in voice mail. Only a small number of commands, such as 'This is financial, don't risk losing bits,' would be needed to tell the telecom company how to treat the data. And the underlying network could be anything from fiber, to satellite, to mobile phone, to cable TV.

As an example, Isenberg cites how the telephone companies might deal with the problem of giving someone an option to make another telephone call while being kept on hold. The telco would research if there was a real customer need, run it through various decision makers, and the request and approve a business case. Plans would then be put in place for development, operations, administration, maintenance and provisioning. Roll-out could take years, even decades. How might the same application be developed for IP telephony? A programmer would develop the option using a standard tool set in the space of a few months, being careful to follow open standards. Then it would be sold from the Web site.

A few telephone companies are investing to involve themselves in the rise of the 'stupid network'. But the majority are very, very cautious. Most money is targeted at speeding up deployment of intelligent network services, says Isenberg, "much like sailing merchants responded to the threat of steam by building faster sailing ships."

For the complete text, see

Date last modified: 24 Jan 98