CTOs Get Priorities Straight: Upgrades, Digital, Data, More


When it comes to finding out how MSOs are doing on new-service rollouts, there's no substitute for going directly to their chief technical officers. Recently, AT & T Broadband's Tony Werner, Cable One Inc.'s Tommy Hill, Adelphia Communications Corp.'s Dan Liberatore and Cox Communications Inc.'s Alex Best got on a conference call with CED editor in chief Roger Brown and Broadband Week contributing editor Fred Dawson. They discussed new-service priorities, equipment availability and other nuts-and-bolts matters. An edited transcript follows:

BW: Tony, how focused are you on rolling out new voice, video or data deployments this year, and which of those options are the highest in priority for you?

Werner: We are focused on rolling out the new services, and I don't think it is fair to put a priority on any of them. We keenly have our eye on probably a half-dozen areas. One is [finishing] network upgrades, because that completes the footprint that we need to enable some of these new services. The other ones are on digital set-tops and rolling them out.

The next one is obviously high-speed data, and then the final one is the phone rollout.

So we are very active with all of those products and pushing them out as fast as we can in the first quarter, and we expect to do nothing but ramp them up over the course of the year.

Hill: We're very heavy into upgrading our systems, and then we're rolling out digital. But we're not into it as much as everybody else [in this conversation]. We are actually firing up our first digital this week. We are going to trial [it] for about four months, and then we will start installing 40 digital headends at the rate of one per week, starting July 1, if everything pans out with our test.

We are also rolling out high-speed data. We have four systems rolled out with high-speed data, and it is going so well for us that we're speeding that process up.

So we will be rolling out probably in the neighborhood of 15, 20 systems before the end of the year with high-speed data. We have no plans for telephony in the near future.

Liberatore: We are continuing on with our [Data Over Cable Service Interface Specification cable-modem] and digital deployments as we have for the last year or so. The DOCSIS deployments are [on] a steady pace. I don't think we'll do any ramping up, but we certainly think we'll have the majority of our homes passed midyear this year.

We do have some outlying systems that we have to figure out how to get interconnected-or [how to] eliminate headends and turn them into hubs.

The digital deployment is virtually the same for us, except that what complicates that a little for us is [the fact that] most of our deployments to date have been on regional fiber rings, where we'd built large multiplexes and then transport those around a regional ring and feed a lot of smaller headends.

Examples of that would be in Vermont, [where there's] a ring around the state. We have a fiber ring that goes from Buffalo [N.Y.] to Pittsburgh and actually picks up Coudersport [Pa.] and some smaller systems through there. That should go over toward the Cleveland area and, from Pittsburgh, come up around through Ohio. So that's complicating it a little for us. We are trying to make sure we understand these networking issues.

We are trying to get [system upgrades] ramped up a little bit. We hope to be done with everything at the end of next year.

Coming [later] this year would be a residential telephony trial and probably an assortment of digital-TV applications, probably midyear this year.

BW: Can you provide a little detail of your telephony and video-on-demand trials, Dan?

Liberatore: Our residential telephony trial will be an IP-based [Internet protocol] service. We believe that's where that technology is going to be fairly soon. So we need to understand how to view that. We are going to go straight to primary-line service [with] network-powered [network-interface units].

We own five switches. So we believe all we need to do is put a gateway in front of that switch, then get the transport from the [cable-modem-termination systems] up to that switch. That greatly simplifies it. But those are our technology decisions.

We did a VOD trial with Diva [Systems Corp.] in Philadelphia. We are going to do another one this year that we hope isn't a trial. We don't do trials. We do deployments, and then they ramp up to multiple sites.

So that's what we believe our VOD effort will be. We will do one site in Pennsylvania and then gradually expand that over the rest of the customer base.

BW: Is there some concern that with the current generation of digital boxes moving so fast and the next-generation, Motorola Broadband Communications Sector DCT-5000-class' box coming right behind it, you will get caught in a generation shift that could strand some equipment or delay service deployments?

Werner: The DCT-1000, 'lowest-level box accounts for a small percentage of our overall deployed base. The majority of our deployed base is DCT-2000s,' which, while they are still not a 5000 in that they don't have DOCSIS modems and a few of the other pieces, they are pretty substantive boxes, certainly capable of doing video-on-demand.

[It's also] capable of doing 256-baud [data], and it is very capable of doing thin-client-based interactivity or a slightly lower level of interactivity. So while we are pushing hard toward the 5000, we feel real comfortable with two or three things.

One is that [the] appetite for expanded programming in digital venues is so high that we recover our costs of these boxes in a very quick time frame.

The second thing is that we will push [earlier boxes] down the food chain, just like we always have, to second sets, in bedrooms or other places, to make way for the 5000.

And we think doing so will [provide] a pretty good, compelling suite of services on an embedded base. We have never had a strategy of slowing down a rollout, waiting for the Pentium III or the Pentium IV or anything like that.

BW: Tommy, you're just getting into this now. You're going to go very quickly, having had the luxury of watching others go through the learning curve. Do you see that 5000-class set-top in the immediate future, and will you hold off to wait for it?

Hill: We would like to have the 5000 box now for selected customers, but that's not possible. But we're not going to slow down our rollout. We are committed to put the 2000 in and, just as Tony said, we feel like we'll have plenty of places for the 2000s that we deploy. And we're thinking that [it would be] probably mid-2001 that we would start deploying the services and the 5000 boxes as needed later on.

We have waited, [but] we are pressing forward-we're going to launch all 256-QAM [quadrature amplitude modulation set-tops]. That's a luxury we had by waiting.

BW: Regarding the DOCSIS process, obviously, we are not yet seeing the start of 1.1 certification. People who are depending on IP telephony as the starting point might be feeling some frustration over that. Or do you feel that you're in pretty safe waters at this point, planning a trial and expecting that 1.1 will be there when you need it?

Liberatore: This is certainly a short trial period for us because we think some of the technology decisions aren't that difficult once you [make certain decisions related to] primary-line [service] versus secondary line, network power and IP.

We'll make sure everything operates well and we can have some confidence in the test plants. Then you have the operational and the marketing trials that are going to happen after that.

To answer your question, DOCSIS 1.1 and the power [consumption] reduction of NIUs clearly are critical items. So I don't feel good about that or confident that we're going to have that. But I'm not so sure there's any way around that for us other than to possibly overengineer and channelize that traffic on its own.

BW: So you feel the opportunity is too compelling to hold back, and you see a way through overengineering, as you put it, to at least begin to exploit this platform?

Liberatore: Right.

Werner: The constant-bit-rate product does work very well, and the costs are coming down. But we also are huge believers in packet-based phone. So while we are doing our initial launches with a constant-bit-rate [service], we are pushing ahead hard with packet, just like Dan is.

You can actually do quite a bit of IP phone even over DOCSIS 1.0, just like [Canadian MSO Le Groupe] Vidéotron [Ltée] is doing. [But] if you want to provide CLASS [custom local-area signaling service] features [like those you get] through a 5E, it's hard [to do] with the soft switches today.

We will be doing a trial later this year where we will put in a device that we call an IPDT. 'That allows you to have IP voice over the cable plant, and then it's got an interface back to the 5E that looks just like a DLC [data-link control] interface does today. So that you can have all of your CLASS features if you've got embedded 5Es.

As far as getting around 1.1, we've got code releases that are essentially a superset of 1.0 or a subset of 1.1, which give you some of the critical parts that you need in the way of fragmentation and other [things] so that you can move ahead with trials in a pretty meaningful way.

BW: Alex, you've had the luxury of not being too concerned about these questions at this point. Although as I gather from your remarks at the recent Voice on the Net conference, you'll gladly jump over when you see the cost parameters and performance parameters you need.

Best: I keep struggling with what will have to occur before I jump over to IP-based telephone service. Our primary objective is providing a lifeline-replacement service. And so we continue to monitor the technology. But the technology has to move to the point where the box on the side of the home draws a small enough amount of power that I can [use] network power, which I do today with the telephone modems I'm buying.

Ultimately, the economics of the overall system have to be so compelling that I'll make the transition. In other words, get the box on the side of the home to draw two or three watts.

Even after that occurs, then either the cost is going to have to be so compelling or there are going to have to be functions and features that are provided over an IP-based phone system that I can't do over the constant bit rate.

At that point, I'll make the transition, but I don't know if that happens at the end of this year, the end of 2001, or exactly when that occurs.

In the meantime, as Tony pointed out, the constant-bit-rate technology has been around now for several years, and it's working quite well. We're hooking up 10,000 new customers per month, so we'll stay with that until there's a compelling reason to switch over.

BW: Talk about some of the specific advanced entertainment services you are looking at beyond straight broadcast digital. There are time-shifted or hosted-recording kinds of things from a centralized headend, or hyperlinking between IP and digital-TV services and the TV to the Web set-top.

Liberatore: We're not exploring it as much as these other guys are going to be, but we're looking at VOD. Right now, our primary focus is to get our digital rolled out, then we're continually looking at the VOD products and which one is best for us, how we want to do it. And we really haven't come to a conclusion on that part.

Where we think there's an important business for us is Web-based TV. We have a real opportunity down the road [if] we can put [data services] on the TV for the people who aren't heavy into computers, so they can do e-mail and surf the Web.

We haven't done any of that yet, but in our future plans, that's where we think this thing could be.

Best: We just announced that we are launching VOD in our San Diego system, using Concurrent [Computer Corp.] technology. So we're very excited about launching video-on-demand.

And as you can imagine, it's not just movies. As an engineer, the kind of delivery platform whereby you can put all types of digital content on a file server in your headend to be accessed by your customer when they want to see it with VCR functionality is a powerful delivery system.

We have had a trial under way in San Diego for a number of months now. After the trial is complete, we will roll it out throughout San Diego and then figure out how to do it across the rest of Cox.

We're still excited about e-commerce. We aren't exactly sure how we play in that space yet, but we think we ought to be able to play in that space. Some folks say [we're] already playing in that space every time someone pays $39.95 per month for cable modems. But we'd like a little bit of a direct transaction fee.

It may be that we play in it with the VOD platform on a local basis as we establish relationships with companies that want to put content on our file server and enable customers to do some type of e-commerce. Maybe that's where we play, as opposed to figuring out how to play with an Amazon.com [Inc.], etcetera.

I find it of interest that all of our stocks have been beaten up a little bit over the past six months or so due to the number of factors. But one of the factors is that the analysts say, "I know you're doing digital voice and data, and that's going great, but that's old news. What are you going to do for me now?"

I think some of the things I just talked about and the other fellows talked about are on the horizon to take us even beyond the road we have to take.

Werner: It's pretty clear that this area is as wide as it is deep, and the power of the DOCSIS channel, combined with the power of the MPEG [Moving Picture Expert Group, opens] a lot of opportunities. [But] we are in a position where we cannot talk on VOD and that type of stuff.

We have done a lot of work on technology with time-shifters, and we think that's just one more tool. It doesn't supplant VOD by any means, but it can augment VOD.

If you've got areas of the country where you want something other than an East and a West [Coast feed], you can put in a server device. It'll take in the feeds and delay them. You can also take and create a bunch of [near-VOD systems].

Of course, a good portion of our focus at this point has been on interactive services. We are putting the pieces together for a fall trial, and it will have some commerce. It'll have information, it'll have some e-mail and some other pieces. You'll start seeing the things burgeoning out this year, and then you'll start to see them picking up some real steam next year.

Liberatore: If you get the right platform out there in terms of set-tops and the right network infrastructure-which we are trying to do with our regional and national interconnects-the next step is to make sure you've got a good strategy for server technology in the headends that allows you to do a lot of these things.

I think once you do that, these applications can start being clicked off fairly easily, and we can really find out what is a good product and what is not.

I don't know about anybody else, but I'm not so sure we know what is a good enhanced-television product. Everybody thinks VOD is the one. But that obviously is becoming something that you're going to have to do because of the competition from people that are putting servers in boxes.

BW: How are you guys attacking the actual deployment issues related to new services in terms of making it more automated? Alex, at one point you said you were hooking up 10,000 consumers per month. If you want to get to millions of cable-modem customers, the process has to get a lot more automated.

Best: I'd say it's still manual-labor-intensive, and let me tell you some of the reasons why. The set-top is the only one of three boxes [set-top, modem and telephone] we're rolling out that is totally under our control from a provisioning standpoint.

We stage them in our warehouse, we determine the authorization levels that they get and we install them in a home. They are totally under our control.

When we install a telephone box, we have to coordinate with the telephone company. We have to get their telephone numbers ported from the phone company to us. The existing phone service has to be disconnected. There is a process we have to go through, and we are beholden to the telephone company that holds us about three days before we could even make that install.

We provision the box and do the installation. But it's an adventure with the phone company. They are not a collaborative partner. It's a similar matter with cable modems.

We are a partner with [Excite@Home Corp.]. @Home actually provisions the service through the CMTS. We roll the truck, do the additional outlet and install the software on the PC. But once again, there are times when we roll a truck, go in and install it, only to find out that there's been a glitch and the service hadn't been provisioned by @Home. We are never going to have these things automated to the degree that we would like to until everything is under our control.

Hill: From the modem standpoint, we're rolling out totally retail. We want the customer to buy them. They're used to buying modems for their computers. We set up dealers in all of our systems where we're rolling the product out, and we decided some time ago not to filter the system, and they're self-installs.

We have our own ISP [Internet-service provider] called Cable-One.net. If [consumers] fire up the modem and go online, they can register and we start billing them. We're not doing 10,000 per day. But so far, it's working really well for us. We really have no one going out to the home and doing anything unless there's a problem.

As far as the digital box, we are working on a self-install deal.

BW: Is there any new thinking in the home-networking situation? How focused are all of you on that issue?

Best: That's a tough one. We are still struggling as a company to figure out how we play in that space or whether we play in that space. We're pretty convinced we play in that space.

But that whole home-networking technology arena is still evolving. There must be 25 companies that are trying to play in that space. I've talked with companies that want to do it over the power lines, the telephone wires, the cable network, wireless, two of the three products, three of the three products, one of the three products. We still have a committee trying to figure out exactly how we sort that out.

In the meantime, we have customers who [want] to interconnect [multiple] computers, and unfortunately, we probably don't satisfy them very well. We will give them multiple IP addresses, then we'll send them to RadioShack to see if they can find a bubble-pack solution for networking their three computers. That probably won't leave them with the best taste.

I'm sure they expect us to roll a truck for free and network their computers for them with some technology we deploy at some nominal fee. I probably think that's where we all want to get to, but I don't think we're there yet.

Hill: We offer more than one IP address for the home and, like Alex said, I guess some time we'll figure if there's a business there for us to do that. But right now, we'll furnish the IP address, and it's up to them to network their own computers.

Liberatore: We haven't seen any driving need as a company to come up with home-networking solutions, and the technology is all over the place. So maybe the technology will converge at the same time we need to have a solution.