|
Dave Nagel is the technological brains behind Michael Armstrong. As president of AT&T Labs, he's responsible for developing both the pedestrian and whiz-bang technology for the company. In this Q&A with Broadband Week editors Matt Stump and Leslie Ellis, Nagel discusses various AT&T broadband platforms strategies: cable, fixed wireless and DSL. His job is part blue sky, part where the rubber meets the road, which is why his colleagues characterize him as "Where the Rubber Meets the Sky."
BBW: AT&T owns cable, DSL and fixed wireless technologies and businesses. When you look at the world, how do you put all those broadband pieces together?
Nagel: We call it Broadband, as opposed to AT&T Cable, because we're interested in access networks that perform better than today's copper pair with traditional electronics. This is one of the key seams and discontinuities in the whole communication network.
If you really look at providing competition in the local markets and significant increases in end-to-end customer performance, the opportunity to create new services, reduce costs, from almost any perspective, broadband technology has an even greater potential impact on the industry than other key technologies like optics, and the core networks.
We see it as a technology that's comprised of a lot of different individual pieces. It would be nice if we could click our fingers and put fiber into everyone's home, but that's not going to happen. It's too expensive and we have to get there in stages. We did a study three or four years ago in which we looked at all the different emerging technologies. We said broadband, in the form of HFC, has the best starting point.
But we also recognize this has to be ubiquitous for it to have the full impact. Where we don't have cable, or can't because of regulation, we have fixed wireless. Where we don't have fixed wireless because that doesn't make sense, we have the option of DSL. And we rank the technologies in that order, but they all have to work together.
They have different performance characteristics. But in the conceptual stack of services, they're all broadband. And we can begin building new and exciting services and better economics in the industry on any or all of them.
BBW: One of the challenges would be stitching up the various backends of the HFC side, the fixed wireless side, and the DSL side.
Nagel: That's true. At the conceptual level, it's just broadband pipes that have different performance characteristics. At the engineering level, obviously there's a lot more to be done. Our first step has been to upgrade our cable plant so that it's two-way and completely digital. We're putting a huge amount of money into it and we're beginning to get very good penetration in markets where we are introducing new services built on that infrastructure.
BBW: Should you have been there by now? Does Wall Street have a beef with you guys in terms of the technology and the deployment of services?
Nagel: I'm not sure I'm the best person to address Wall Street's beefs. But we've done pretty much what we said we were going to do. The percentage of the network that's upgraded to two-way digital is pretty much where we said it was going to be. We've started rolling out telephony services. We said last year was a year of engineering investigation and market trials, which we have done.
We've got nearly a million high-speed data customers, or will have by the end of the year, and we're going to make the numbers pretty close on telephony. I think where Wall Street frankly has been more disappointed is in our traditional services.
From a technologist's point of view, there have been no major bumps in the road. If someone says it's not as fast as we thought it was going to be, well, I'm sorry that's the case. I think it's going very, very well and customers love the services, which is the most important criteria.
BBW: In your phasing process where you rank HFC first, then fixed wireless, then DSL, what does that mean in context of the Time Warner telephony discussions? Could it potentially happen that if you are not able to strike a deal that is reasonable for both of you that you would offer a fixed wireless bundled service plan in AOL Time Warner markets?
Nagel: HFC is the most mature. DSL has more homes passed but not all of the central offices are equipped and obviously that's not rolling out as fast as some of the early DSL supporters said it would.
So we still like HFC. I still think it would be a great partnership to work with Time Warner, but ultimately we're going to be competitive. We see the nation as our market. If we can't cut a deal with anyone, we're going to compete like hell in those local markets with whatever the best technology is.
We can deliver superior telephony over any of the existing broadband technologies. Video would be a more difficult challenge over something like DSL or fixed wireless, but we can take telephony anywhere and do a great job with it. And we will use whatever networks we have available.
BBW: Last year, you were installing fixed wireless in Dallas for $800 a home, with the hope that costs would come down over time. How is the price equation coming in for fixed wireless and what's the status of Dallas?
Nagel: The price is coming down just as we said. I don't have the exact figures but it's certainly south of $800 already. We only have a few thousand installations, so its pretty early in the life cycle.
Generally, the equipment is going to follow Moore's law. The price is going to come down pretty much at that rate. Maybe a little less fast in the early deployment, but over the next few years I expect it to get there. Most of the system economics are heavily dependent on the cost of the consumer premise equipment at the network's edge. We're doing extremely well in our fixed wireless trials. My primary metric is what the customer is saying. All the data suggests they love it and think it's very, very high quality. They consider it to be the equal of, performance-wise and reliability-wise, with what they've had for many years plus all the advantages of having a broadband data connection.
BBW: What's the data throughput rate?
Nagel: Initially it was 512 kbps, and it will go to one megabit. I think they are still operating at 512. It's still more of an operational decision than a technology decision and we will bump it up to the megabit range certainly in the next year or so.
BBW: What are the plans beyond Dallas?
Nagel: We've launched in Dallas and Fort Worth and have 6,000 subscribers. We plan to launch in San Diego, Los Angeles, Houston and Anchorage by year's end and have 40 markets online by the end of 2002.
BBW: You alluded to integrating HFC, fixed wireless and DSL plant and back office capabilities, I believe. When do you have to start worrying about that?
Nagel: It's important that we be able to provide broadband service to any of our customers anywhere. If you look at where we have spectrum licenses and can operate fixed wireless, where we have cable properties, and where we can get access to the rest of the people through DSL, from a deployment point of view it's patchwork.
But it does come together certainly at the level of the IP backbone and the IP network. That is the point at which we can begin to provide common services, common infrastructure, common operating systems over the broadband networks independent of whatever physical access network is used.
So my group focuses on developing communication services, messaging, telephony, etc., which work independent of the physical mediums. They are quite happy to work with a bunch of servers that run in the middle of the network. And they are quite happy to serve up DSL, or fixed wireless, or cable, or in some cases even narrowband.
BBW: Where are you on IP telephony?
Nagel: We started with circuit telephony that we could buy off the shelf and operate at very high levels of performance and reliability. It's done very well. There are strong financial incentives to go to packet. Packet unit costs or telephony costs get better as the cost of the customer premise equipment comes down.
You're looking at significant potential savings in capital equipment if you can get as fast as possible to a packet solution. The biggest impediment is that the vendors can't provide today all of the equipment and software needed to provide seamless interchange with the circuit stuff. They're beginning to get there in reliability. They're not quite there in features yet.
My hope and expectation is that they will be in the next 12 to 18 months. We're very aggressively developing all the infrastructure we need to take the vendor's equipment and integrate it into the network. We have a team of 30 people working with the broadband folks to get this stuff into the marketplace as quickly as we can.
And most exciting to me, given my gear-head status in the company, is the ability to integrate new communications services that go beyond plain old telephony, such as various sorts of unified messaging. We have shelves of really fantastic innovative services that we've developed in anticipation of broadband deployment. I'll just mention one, "voice tone." It replaces the old dial tone that we are all familiar with, with a voice, that says "How can I help you?"
BBW: AT&T Annie!
Nagel: Yes, that sort of thing.
BBW: How do you handle the technology migration from circuit switch to IP?
Nagel: The particular approach we're taking is to get it in the home first because that's the hard part, both from an operational and logistics point of view, as well as getting the equipment ready. Given that the manufacturers aren't ready with soft switches, what we've done is a very clever solution that basically makes it IP all the way up to the headend. Then there's a device that translates it from the packet world into the circuit world. Then we use the whole 5E infrastructure which we already have. When we get soft switch technology, it becomes plug compatible with the 5E. We just swap out the 5E's and begin to introduce the digital soft switching without having to miss a beat.
BBW: What's that device that translates the IP packet at the headend to the circuit switch world?
Nagel: It's called an IP digital terminal (IPDT) or network call signaling gateway that converts from IP to GR303 for the switch interface. It basically takes the packet and strips the headers off and puts them back into a streaming format and into the standard circuit based telephony standards.
End-to-end IP requires a server, or call agent, that has call processing capabilities similar to a switch. Calls that can't travel end to end on an IP network go off-net to the voice network through a voice gateway.
BBW: Right now you have three $200 to $300 boxes in the home-the digital set top, the cable modem, and perhaps a broadband telephone interface. Do you keep those three devices separate or is there a need for one box?
Nagel: There will continue to be separate boxes because not everyone is a TV subscriber that wants to be a phone subscriber, etc. If you try to do the grand integration scheme, you would find the economics wouldn't work very well for customers who took only one or two services. On the other hand, from a box designer's point of view, you can share power supplies and boxes, and other pieces that for someone who takes all of them, it's a much lower cost. So there probably will be fully integrated solutions, hopefully when the retail channel begins to get cranked up.
BBW: Let's go to Excite@Home. It's part content service, and part backbone. How do you see it gestating over the next two years inside AT&T? How does it integrate with the rest of your technology plays?
Nagel: The original idea of Excite@Home was a platform for broadband services. They got out to a head start and still have a lead in that. It's a great approach, tying all these different broadband access networks together and providing a seamless platform for people like myself and others to build services on top. We still really like the Excite@Home model a lot.
BBW: But isn't their backbone essentially your backbone, and aren't there those who would criticize Excite@Home's excessive spending on content versus the network?
Nagel: Well, those are two different questions. They have a state-of-the-art, very high performing network. Our backbone is very high performing. It's OC-48 all the way around. They have a state-of-the-art server complex. We've yet to have to split a node.
BBW: Aren't those OC-48's AT&T's?
Nagel: In effect they are. We provide them in a fairly raw form and the @Home guys have stitched them together into a network. I guess the most accurate way of saying is that we provide the links, they have created a network out of it. So it's very much their network.
Is it a good idea to link the content in an intimate way to the broadband network? That's really predicated on what I think is a good idea, being that broadband is such a different qualitative step in performance. You could develop specific new content that would take advantage of that broadband network and accelerate the overall movement toward broadband. If broadband just means that you do what you do today just a little bit faster, it's less interesting than if you could do completely new things. The whole idea about putting Excite and @Home together was to accelerate the overall deployment of broadband.
Whether the best way to do it structurally is to have the companies joined is something that is left to people other than myself. We at AT&T see less value in that type of business integration. We think broadband content and new applications can be developed independent of the broadband platform and in fact may be done best separately.
BBW: What's the status of rolling the LightWire architecture into other systems?
Nagel: It's been rolled out in Salt Lake City, and we have found that it works extremely well. We've learned that you have to deploy it very carefully from a network planning point of view if you don't want to spend too much on the equipment. In other words, the network design has to be optimized carefully from a topographic point of view. But we continue to think it's a promising technology improvement and is much better than the standard node splitting techniques that have been used for a long time in the cable industry.
It's very promising in terms of increasing the overall performance and reliability of the network. Since we replace active components with passive and more reliable components, we can get much better levels of reliability in the network. At the same time, we can reduce the operating costs because you don't have to be tweaking amplifiers as much.
BBW: Is it ready for prime time or do you wait until IP arrives?
Nagel: I'd characterize it as ready for prime time. You know, I'm on the technology side so I'm probably always a little more over enthusiastic about this stuff than some of the operational guys. In terms of actual deployment plans, that will really depend on where we are in the network build and where it falls in the priority of how to spend money to execute against our business needs.
The best way of thinking about it is that it's something we have on the shelf. And as we go from initial upgrades to wanting greater reliability and greater performance, we have a ready-made solution.
BBW: What are you doing with DSL in terms of deployment? And talk about voice over DSL, video over DSL. How scary are they for you as a competitor?
Nagel: I don't think they're scary at all. First of all, despite all the scurrilous ads that are out there about sharing your bandwidth, cable is a better solution. In areas where you can't get cable, we think it will be good for telephony, and it's good in a limited degree for data. But it does have some disadvantages, not the least of which is that the people who own a significant part of the physical infrastructure that you need to deploy it are the local telephone monopolies. And they have shown themselves not surprisingly reluctant to deploy it in large quantities.
The problem the locals have with this is it craters their high profit data business because it uses the same technology, it just sells it for a lot less. That's something a lot of people don't like to do. The problem, which Covad and Rhythms are finding out and is reflected in their stock prices, is that working with the local telephone monopolies to get this stuff deployed is not a particularly pleasant or efficient thing to do. The operational complexities of reselling someone else's physical network is tough. That's why it's third on our list.
BBW: There is some research that as DSL penetrations rise, crosstalk is a time bomb waiting to go off. Is that your group that did that research?
Nagel: We have done research like that. You're absolutely on target. This idea that anyone is going to get point-to-point all the way from end-to-end access to a single router or a single port on the network is nuts. Of course it's shared. It's all shared. It's just a question of where it's shared. I'll just repeat what I said earlier: We've never had to split a node because we ran out of performance on our high-speed data networks.
BBW: And they don't have nodes they can split.
Nagel: And they don't have nodes they can split, that's exactly right. So this is the kind of marketing that appears when you have an inferior product.
BBW: What about SBC's Project Pronto? They're moving gateways deeper into the system.
Nagel: I probably don't know enough about the details of Project Pronto to comment intelligently about it. From what I've read, I think it's more of a proof of concept and a demonstrator than it is something that has any practical significance in the marketplace.
BBW: What's juicing you these days?
Nagel: What's fun is that we're just now at the point where we can deploy services, and get this stuff in the hands of customers. It's going to be a very exciting next couple of years as all sorts of new services evolve-from taking advantage of mobile broadband networks that are being built in the next couple of years to the broadband fixed networks that we already are deploying in increasingly large numbers and with commercially interesting numbers.
You're going to see all sorts of great new services. My favorites are the streaming services- audio and video. It's going to be a wave of innovation that's going to be really terrific for customers and good for the industry.
BBW: Can you see 3G wireless wedging itself in above fixed wireless, for instance, in that tier that you talked about earlier?
Nagel: It's a couple years behind fixed wireless. They will probably maintain that separation. Fixed wireless will get better and faster in future versions and mobile will probably continue to be a couple of years behind.
But even with the early, GPRS levels of performance, at 100+ kilobits per second, you're going to see some really exciting, innovative new applications that take advantage not only of the speeds, but the always-on characteristic. You will be able to consider it just a mobile extension of the Internet.
We're developing some really nifty new service concepts, location-based and otherwise, that take advantage of that. The streaming media stuff is going to be really important. In the mobility arena, it's pretty clear that it's going to be audio first, but even some limited graphic or video stuff will probably come in the next year or two.
|
