10 minutes with Vinton Cerf

In a chaotic and noisy lobby outside the conference hall where he was just awarded the William G. McGowan award for leadership in overcoming obstacles to global connectivity, WorldCom's Vinton Cerf sat down for a brief interview with Telephony's Tim McElligott. His stream-of-consciousness remarks to the last question indicates that the Father of the Internet is still like a kid in a candy store.

Where does most of the work need to be done for VoIP?

The most interesting work will come as we see new services for the SIP servers and IP-based devices. And I use the term "devices" on purpose because the term "phone" is such a narrow notion. You and I both know that the devices we carry today are more than just telephones. They have IP capabilities and browsers and short message capabilities. So I see the most interesting work in the expansion of services based on SIP. The most difficult work, though, will be getting more experience with the full range of vendors [currently] developing software and hardware. This is still the early days in the deployment of VoIP. We need to get a better sense for what vendors are capable of doing, whether they interwork as people make IP calls from one device to another, whether they use the same compression algorithms or negotiate properly to be able to find each other in the network and add a third or fourth person. There is a lot of vendor interoperability work to be done and discussions about network management that need to be sorted out.

Given the time it took for early developments in IP, are people not being patient enough with development of the VoIP industry?

No. Patience is one thing; persistence is another. I want to make that distinction. It is inadequate to be merely patient; one has to be persistent. You have to keep pushing because if you stop pushing, sometimes nothing happens. And you have to realize that you may have to push for a while before something actually gets done. It has been 30 years in the making. We're only just beginning to see the fruits of voice services on the Internet backbone even though we were doing tests of this capability in 1975. The quality wasn't as good, and you could only carry about two voice conversations over the backbone before you ran out of capacity, so a great deal has had to happen. The thing is, part of it doesn't happen until conditions are right. Sometimes things don't get invented until it is possible to invent them, and then it's like simultaneous invention where more than one person does it. In the case of IP telephony, it didn't really happen until we had enough computing power and bandwidth, and low enough delay to make it work better. Then around 1996, it suddenly blossomed because people had laptops and high enough access speed to make it work better.

You propose a central database for ENUM. Is there any benefit to a relationship between an ENUM database and location databases?

That's a very interesting question. I think geographical index databases are going to be incredibly valuable as more and more devices [can] know where they are. I don't immediately see much connection between the ENUM-style database and the geographically indexed database, but I do see the immense potential for having geographical index data that I can get to through the Internet in the context of knowing where I am.

So if I am in, say, an Internet-enabled car and I'm talking on a cell phone to a computer that understands what I am saying, and that cell phone is part of the local net of the car, and it has learned about the Internet address of other devices in the car, then it knows about the GPS satellite computer that is in the car and it can interrogate the computer to find out where it is, and while I am talking to this computer that is somewhere out on the net--whether my voice is carried through the net or not is irrelevant--if I ask it where the nearest Thai restaurant is--which is a reasonable question to ask--the computer looks up in the geographical database where the restaurant is. But how does it know where I am?

Well, the cell phone got that information from the local Net and passed it along as a piece of data and what comes back to my ear is that it is two blocks away on the right. And what pops up on the navigational display--which has its own Internet address that is also passed to the computer listening to my voice--is a prompt asking if I want to see a menu. Then it reaches out to the Web server of the restaurant and pops that up on the navigational display and asks if I would like to place an order. What's happened, and the thing that gets me all turned on here, is that a variety of Internet-enabled devices are being drawn together temporarily in an ensemble in order to service a request. It isn't just client/server anymore. It's this rich fabric of interaction among devices that can communicate with each other.