Last Modified: ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?



Homepage / Publications & Opinion / Archive / Interviews (pre 2002)

Fortune Teller
tele.com, Issue: 412, June 7 1999

Fortune Teller -- Talk about future shock. BT's Peter Cochrane is taking technology where no service provider has gone before.

Meg McGinity & Carl Weinschenk
A world that can only be imagined by most is clearly a reality for Peter Cochrane. But glimpsing this new world through his eyes is not for the timid or the technically conservative. And it's not for those interested in a quick intellectual dip. Cochrane's hushed, matter-of-fact tone and British accent may momentarily cushion the full impact of his mind-numbingly revolutionary predictions, but his ideas resonate hours, even days, after any conversation with him has ended. Fortunately for Cochrane, his visionary bent suits his vocation well. As BT's chief technologist and director of the innovative BT Labs, Cochrane makes it his business to determine what will turn into tomorrow's service. And some of what he envisions can be unnerving. Cochrane, who has more degrees than a thermometer, believes the future involves a preponderance of not human-to-human but rather machine-to-machine communication. It's a world that will be filled with mobile wireless connectivity, where picocells that cover small defined areas will be a prerequisite for supporting the vast increase in wireless usage. It's also an environment where service providers will have to invest heavily in third-generation (3G) wireless technology-an important advance they can't afford to ignore.

The changes he sees coming don't end with wireless. Cochrane says that by 2010, the biggest provider networks will carry traffic that is 95 percent generated by computers and only 5 percent by humans. These visions represent major challenges for Cochrane. As the driver of technology development at BT, the United Kingdom's biggest service provider and AT&T's partner, it's his job to explore future technologies, systems networks and services, and to flag BT on where it needs to be looking to keep pace with competitors.

In Cochrane's vision, these inevitable changes will blur the lines between man and machine even further, especially as applications such as silicon chips become commonplace. These chips-offered perhaps by service providers-will be embedded under a person's skin and will house information on bank accounts, health profiles and even passports. They might also provide more than personal data: They could be used to trace a person's whereabouts, which would ease the concerns of parents or kidnap-phobic executives. When you hear Cochrane explain it, you think "the tagging of people," as he calls it, isn't so preposterous. "With the abundance of foreign objects like pacemakers and breast implants, technology has already found its way into the human body," he explains logically when faced with shocked expressions. Artificial life, artificial intelligence and wearable computers: They're all part of Cochrane's tomorrow. And, he believes, everyone else's.

Senior editor Meg McGinity and executive technology editor Carl Weinschenk recently spoke with Cochrane at the BT offices in New York.

Are there any technologies or developments that service providers are missing the boat on by not yet investing? There are a couple of features that are starting to scream out there. One is third-generation mobile. Another is machine-to-machine communication. There are about 14 billion microprocessors and 6 billion of us, and increasingly machines will want to talk and communicate. About 40 percent of all the traffic my company carries is machine to machine, not person to person. We have a gross curve-it says that by 2010, 95 percent of the traffic will be machines talking and only 5 percent will be people. So we will have to start thinking about some fairly innovative changes.

What are your predictions about what we'll need for the changes? Right now, the cellular mobile phone network has cells that are typically 10 to 15 kilometers across. If you get into a city, they may be a couple of kilometers across. We're going to need cells that are 200 and 100 and 10 meters across. We're going to have to put up both microcells and picocells. Instead of having a network that has a lot of very thin parts that lead to bigger parts, it's going to lead to a network that's fat everywhere.

Is this suggesting a network that's passive in the middle? Yes. The problem is mind-set. There are still people in networks who think that distance and bandwidth are important-but they're both irrelevant. Distance and bandwidth don't cost anything. What we are looking for is very low-cost, very fast access. It doesn't cost a great deal to do that. I could show you pictures of a cellular radio system that's smaller than this tape recorder. We have the ability to put a cellular site up in every room, in every building or on every street corner. Then you could provide bandwidth everywhere very easily. You can daisy-chain these sites so you don't actually need any intelligence at all.

Why then hasn't wireless data taken off? Let's have a look at some cost numbers. I have a radio system in my home. The distance from the tower of the BT Lab to my home is just over a mile. I've got a T1 radio link into the attic of my home. I've got a T1 system that beams down T1 into every room. Eighteen months ago, that system cost $10,000 for the box. I've just replaced that with a $700 box. Now that's starting to get to the right number. If you can get that to be a $300 box, then the sheer convenience says to you it won't be the network provider who does that, and you'll be prepared to buy your own box. You don't have to have the guys digging up the garden or stringing cable up.

A lot of infrastructure is in place. In the U.K. we have installed four million kilometers; we've got more fiber per capita than most nations. The Internet statistic is that 85 percent of our customers are all within 200 meters of our fiber. And we have a slight impediment: Government legislation won't let us use radio right now, so we may have to go into a sort of partnership and that kind of stuff. But there are not-as far as I can see-any technological problems in doing any of this.

You've described some paradigm shifts. How will we get there from here? I think there are some very interesting models. Take satellite broadcasting in the U.K. They don't provide the set-top box. You do. But the set-top box is subsidized. You get it at less than cost. This is like the laser jet printer. You buy a laser jet printer for $150, but the ink refills cost you $80 a shot. They get you on the refills and the paper. In my model, the PC or laptop becomes a $300 commodity item. You'd get everything for almost free, then they'd get you later on the service, the advertising, the stuff you buy later.

This is happening with phone service in the United States. You could pay not to listen to individual advertisements on your phone. Or you could pay a monthly fee to switch all the ads off on your phone. Then everybody could have a phone, but if they wanted it for free, they'd have to listen to the ads. The battle is for the eyes and ears.

So is the migration of circuit- to packet-switched networks a baby step toward all this? Let me be slightly contentious. There's not a hope in hell that IP [Internet protocol] or ATM [asynchronous transfer mode] will replace and provide the service quality that you get in a phone network. It will never happen. Are you saying a lot of people are spending bad money? Here's what's going to save their butts: The key parameter that we should worry about when we talk about communications is delay. What's bad on the IP and ATM networks is the management of delay. What you actually need is circuit switching, no delay. What's the solution in IP and ATM? They're going to bolt down the routing path so that you get constant delay.

So there are several things that have to happen. One, you've got to reduce the number of hops. The reason: If there's a probability of blocking some node and you can catenate a whole bunch of nodes, you don't need a whole lot of nodes before you get virtually no bits getting through. What you want is to have the packets go through the same routing every time it's bolted down. That's actually circuit switching. So the IP and ATM community have just reinvented circuit switching.

Can it work? For it to work, you need to displace somebody else so that our bits can go through slick. Who are you going to displace? If you start, what are we going to do? Will there be arm wrestling, as in "We're more important than you" or "We'll knock you out of the way and make you wait, but we'll pay more"? So we could pay more for less delay. You could have grades of service, and if you've ever made an IP call, you know how bad it can be. If you've ever made a call using a satellite phone network, you know how bad it can be. It's absolute garbage. What's going to save it is 95 percent of the traffic being machines, because what we will do forces them to wait while we come straight through.

Once you segregate the machine traffic from the voice traffic, can't you start using the old contention schemes that telephone companies use to predict how much traffic will be on their networks at any given time so they can deploy resources? Yes, you can. So I think it's going to work. But if it were all human beings, it'd be a hell of a scrap, wouldn't it? Because what you're going to have to do is overprovide capacity to just make sure that you and I don't get any delay 99 percent of the time when we talk. For person-to-machine communication, is speech recognition ready for that yet? I think it's getting pretty damn good.

What would you say are the service providers' top three problems that must be solved? The first one is the legacy problem. My company has a network investment that's around $42 billion. We invest in that network in the U.K. at around $4.5 to $5 billion per year. Now, if you make a mistake, you feel it 10 years later. The technology's moving so fast that you might well invest in a whole bunch of stuff, get it installed and actually not put it into use before you have to rip it out and replace it. We've come from an industry that said you install the switch and you depreciate it over 25 years and it might actually last 40 years. Now we're at the point where you install some ATM stuff and then say, "Hey, we decided we're not going to do that anymore." You rip it out, even loaded with traffic. This is a big problem for the mind-set. One of the key problems for the network providers is picking the right technology and topology for the network, then getting the right services in there that allow you to make enough money. These business models are absolutely key to survival. There have been an awful lot of experiments of very expensive things that we know people won't pay for.

What is the most far-out application that you are working on right now? The one that you think would sound like pure sci-fi if we discussed it without you in the room? Oh, I think the tagging of people. You mean having a chip implanted under your skin? It seems far-fetched, but let's look at it this way. There are now hundreds of thousands of people with pacemakers and electronic devices embedded in their bodies. So let's say you're walking down the street and your pacemaker knows you're going to die in 10 minutes and it makes a phone call. It just seems to me that there's going to be a bunch of electronics inside people that's going to want to communicate. Or take your passport. Somebody looks at this photograph and says, "Yes, that's you," and lets you in. I just want to wave my hand and walk in. I don't want to carry any cards. I don't want to carry any money. I found out the other day that some Italian diplomats are having chips embedded in their skin to guard against kidnapping. It's happening.

What about the threat of a bandwidth glut? Well my kids have got one T1, and they complain bitterly: "When are we going to get some serious bandwidth in here? This is crap." We went from 28.8 kbit/s to 128 kbit/s to 1.5 Mbit/s and all this and, gee, it's slow. When is it going to get really fast? Because at work I've got 100 Gbit/s around the lab, I've got 100 Mbit/s plus at my desk and I've got 100 Mbit/s around my home-so machine-to-machine is lightning fast. But when you go on the Net, it's World Wide Wait ... tick ... tick ... tick. Ultimately, the limit to your ability to produce is the Net. That response time is absolutely vital.

What will be the major societal changes from the Internet? The newspapers are not going away, and we've got more magazines now as a result of electronic publishing. My kids play rugby, swing from trees, play chess, go swimming, ride bikes, go fly fishing-and do their homework on the Net. I've watched it change, though. Kids used to go from book to CD to Net. Then it was CD, Net, book. Now it's Net, CD, book. They go to the Net first and invariably find what they want.

So, do your kids read as much as you did when you were young? That's an unfair question because there were only three books in my home. We were a very poor family, and so I went to school to read. But the answer is, yes, they read. However, reading can be an awful waste of time. I think what is key is the acquisition of knowledge and experience and understanding. I don't care what the mechanism is. But I know they're better educated than I was, for one thing. They know more than I did, and they understand a lot more.

Carl Weinschenk can be reached at .
Meg McGinity can be reached at .
Copyright (r) 1999 CMP Media Inc.

All materials created by Peter Cochrane and presented within this site are copyright ? Peter Cochrane - but this is an open resource - and you are invited to make as many downloads as you wish provided you use in a reputable manner. Website maintained, hosted and designed by USP Networks Events Review