Because they are crossing swords with multiple enemies on multiple fronts, cable operators can no longer afford to concentrate on just one new service at a time, but must be able to launch and support many as they act on their own and react to the EchoStars, DirecTVs and Verizons of the world.

CED Editor Jeff Baumgartner and CED Contributing Editor Leslie Ellis recently caught up with some of cable's top technology talent to learn more about cable's current stable of weapons, as well as those that are still in the foundry, being built to help the industry meet the challenges of the not-so-distant future.

Joining the call were David Fellows, the EVP and chief technology officer of Comcast Corp.; Marwan Fawaz, the SVP and CTO of Adelphia Communications; Mike LaJoie, the EVP and CTO of Time Warner Cable; and Wayne Davis, the EVP of engineering and CTO of Charter Communications. CED held a separate conversation with Chris Bowick, the SVP of engineering and CTO of Cox Communications. An edited transcript follows.

CED: Let's start with digital simulcasts. Where are each of you with deployments and trials?

Bowick: As far as our plans for this year are concerned, we will have digital simulcast up and running in some shape, form or fashion—and this doesn't necessarily mean all of the analog programs will be digitally simulcast—in about four markets this year. The anticipation is that we will move much faster in 2006 and beyond, but we're starting off in 2005 with probably about four.

LaJoie: We have some trials up now. We have plans to roll it out pretty much everywhere, as fast as we can, as part of a larger bandwidth initiative. We won't get it all done this year.

Fawaz: We're in trials. We have plans for launching our top 10 markets this year. The goal is similar to Mike's: We'd like to set the foundations for the transition from analog to digital, which obviously will take a long time. You have to start somewhere. In conjunction with simulcast, we are employing additional spectrum efficiency tools, like advanced statmuxing. We're also looking very closely at switched digital.

CED: How many channels in the mux?

Fawaz: Today we have a lot of SD/HD (standard definition/high definition) combinations. So we do two HDs and one SD (per 6 MHz channel), which helps us to take advantage of some of the leftover capacity on the HD QAMs. And we're also going from 10:1 to 12:1 of the multiplex.

David Fellows
‘Analog is going to stay around for quite a while– and a lot of analog. Sixty channels or so. The purpose of simulcast is not to reclaim analog channels on day one
But it sets up that possibility.’ –Fellows
Fellows: We've said that by the end of this calendar year, we'll be simulcast everywhere. We're already launched in a few systems, where digital boxes now tune to the simulcast channels. The only reason we aren't going faster is the availability of the ad insertion gear. We insert ads on something like 40 different channels. We make more than a billion dollars a year doing it, so it's important to have that as part of our simulcast effort. We have gear in our lab that does GigE in, GigE out. But more of our systems haven't switched to the simulcast because we don't have (ad insertion) deployed in the field yet.

Davis: We're obviously commercially launched with simulcast in Long Beach (Calif.). Now we're expanding outside of the Long Beach market, into the larger L.A. footprint. As well, we're in the process of rolling simulcast out over the next 30 to 45 days in our Wisconsin markets. It's active and turned up in the headend, and we have friendlies on it in Wisconsin now. There are other markets behind it—which I won't speak about now—but, in general, we're looking to pass 3 million homes with digital simulcast this year. A little challenge we have is the number of headends, and the cost. So there are transport and connectivity issues now. As far as the bandwidth goes, the additional bandwidth required for this—we're taking a very traditional approach. We're moving to 100 percent 256 QAM where we're not already.

CED: Chris, you mentioned that you might not simulcast every analog channel. What factors would keep you from simulcasting some and not others?

Bowick: Remember that this is a migration over time. I think a lot of our field locations are just now beginning to play with the concept. Part of it is that. Part of it is making room for the additional 42 to 48 MHz of additional bandwidth in order to transmit another 70 channels, so there's some work that has to happen there. Just miscellaneous stuff—like ensuring that all of your analog pays and pay-per-views are up in the digital tiers as opposed to down in the analog area, and making the transition to 256 QAM across all of your network to ensure you've got the 40 percent efficiency gain there. Also, we're looking at ways to improve multiplexing efficiency—moving beyond the 8:1 for standard definition to something above 12:1 or even beyond to gain some efficiency there so you can make room for some of these simulcast networks.

CED: There's been talk for a couple of years about this $50 all-digital set-top. How is that coming, and what feature sets are you currently expecting that device to provide you at the get-go?

Bowick: We're actually in the process of working out the answer for that. We do have a concept for a low-cost box that we've proposed to both of our primary vendors. We're also aware of their low-end boxes—the DCT700 on the Motorola side and the Scientific-Atlanta equivalent box. I think in general the feature set that's available in those types of boxes is palatable to us, but...our transition to OpenCable and OCAP (Open Cable Application Platform) is very important to us, and, of course, we've been pushing OnRamp to OCAP, so we'd certainly want to ensure that any boxes that we deploy would be capable of deploying OnRamp to OCAP as a minimum and potentially a full OCAP later on.

Frankly, we are paying very close attention to the work Comcast is doing with Motorola and the work that Time Warner is doing with Scientific-Atlanta. We obviously have to face the fact that those two do have a tendency, because of their volumes, to drive those two vendors pretty significantly. I don't think you'll see our needs vary significantly from the rest of the industry in that regard.

Fellows: At Comcast, we have an active program to both produce the chips for that box, and to have vendors produce that box. It has DSG (DOCSIS Set-top Gateway) for signaling. It has an advanced video codec, either VC-1 or MPEG-4. It's capable of running OCAP, so that dictates a certain memory footprint size and processor speed. And it is all-digital, so, no analog (components) are necessary. That helps you save some money. The box next year will not be $50. In the same way, cable modems weren't $50 to start with. We're trying to make sure we have the right $50 DNA inside the set-top box, as it will get to $50 when we crank up the volume.

LaJoie: I'm not comfortable in an environment where there [are] multiple MSOs talking about the prices we pay. It's not a conversation we typically have. We all drive toward it. But we don't talk about price, for obvious reasons. Having said that, I agree with Dave: There are ways to cost-reduce the box. I think the cable modem analogy is a good one. The steps that the industry is taking will have a similar effect on set-top boxes. There are similar kinds of cost curves. The set-top box is still more complicated (than a cable modem); there's more circuitry. Besides, if you really want to get into the expense of the box, you can take the clock off the front, take the buttons off, and make it a very smart, very simple network device.

Davis: It is important for that (set-top) price to go down. As we switch to all-digital and put it in front of all the analog devices out there—having that box be less expensive is very important. It's a hurdle to overcome. We need to drive it down.

Fellows: I forgot to mention, in my list of attributes, one important addition to the things we want in that low-end box: Downloadable security.

CED: Are you developing that yourself?

Fellows: It's one of the CableLabs NGNA (Next Generation Network Architecture) projects, so it's being worked on under the CableLabs umbrella. But yet, chips are being developed, and software is being developed. Downloadable versions of today's CA are being worked. (Editor's note: Soon after this conversation, Comcast and Motorola inked a far-reaching partnership that included two joint ventures concerning Motorola's conditional access technology. Please see p. 14 for more details.)

Fawaz: I'd like to follow up on the feature list Dave just mentioned. Obviously, an all-digital box has to support all incumbent apps—IPG (interactive program guide), VOD (video-on-demand), and so on. So I think there are two main questions that are up for debate. One is the advanced codec, which is a unique opportunity to help solve some of the bandwidth issues. The second is whether we want it to support OCAP or not. That's also debatable. Obviously you're looking at a higher memory footprint—maybe 16×64 Megabyte—that adds cost. So we're taking the analog contents out, but we're adding things that take the price back up. It's a tough balance.

CED: Does that mean you'll have a range of devices, even on the simple, digital-to-analog, low-end side of things?

Fawaz: You'll hear suppliers talking features. They challenge us. Do you want DOCSIS? Adelphia's answer is yes. As for how much processing power and memory we need to support OCAP... the business case needs to be run.

Mike LaJoie
‘You’re not going to see Time Warner expand spectrum above 750 MHz. There might be an isolated case here and there where we’d decide to do it—but by and large, no.’ –LaJoie, on bandwidth expansion
LaJoie: I don't believe we'll be buying boxes down the road that don't support OCAP. The boxes we buy will support it. I think that the real dollar floor on set-tops is going to drop, without a doubt, even taking into account all of the features Dave outlined. There's not one that he said that I wouldn't agree needs to be there.

You'll also see us deploying more expensive devices as well. It'll mostly be dictated by market conditions. And, we're going to see a lot of devices connected to our network that aren't owned by the cable industry. The two-way CE (consumer electronics) devices that are coming—they'll take some cost off of our balance sheet.

CED: What about analog spectrum reclamation? When does that start to kick in?

Fellows: Analog is going to stay around for quite a while—and a lot of analog. Sixty channels or so. The purpose of simulcast is not to reclaim analog channels on day one. But it sets up that possibility.

Fawaz: Although it's a small part of the industry, there are still some limited bandwidth systems out there, with less than 750 MHz. The question is, do we want to upgrade those by going to 100 percent digital? Our position has been, we don't think you want to make that 100 percent cutover overnight. We don't think it's consumer-friendly. I agree with Dave. You'll have a lifeline analog service for a while. It will be a gradual transition, using a simulcast approach. We want to make the move in a consumer-friendly way.

CED: Despite the shift to digital, do you think analog can retain its value long term? I realize that you're early on with this, but do you anticipate offering a lifeline analog service?

Chris Bowick
‘The technology is, I believe, going to be an important one for all of us to embrace at some point in the future, because there are significant savings to be had from a bandwidth perspective.’ –Bowick, on switched video broadcast technology
Bowick: That's, honestly, a very debatable question both in the industry and even within Cox, because it's a very strategic discussion. Many have viewed, and I fall into this camp, that analog certainly takes up a lot of bandwidth— there's no doubt about that. And we need to be more efficient in how we use that bandwidth. And yet when you look at the strategic advantage that analog carriage has over our competitors, especially over DBS, for example—we can place A/Os as many places as we'd like without having to place a box on these bedroom and backroom television sets. To that extent, it's certainly a strategic advantage, but there's probably going to have to be some improved way of utilizing that bandwidth, because when you look at everything we'd like to add in the future and try to cram it all into 750 MHz of bandwidth, it's going to require over time that we use some advanced technologies. But in addition to that, it might also require that we recapture a little bit, at least, of that analog programming. We've looked at scenarios internally that could take us down to as few as 20 analog channels or, on the flip side, keeping all of our analog channels.

We've looked at both scenarios. Both are doable with aggressive use of advanced technologies from a compression and multiplexing standpoint and from a switched video perspective. In my estimation, what you would see longer-term is something in between.

CED: Are any of you with systems below 750 MHz thinking of doing a cutover without the simulcast?

Fellows: I'm watching with great interest the parts of Wayne's systems that have gone all-digital, just to see what kind of surprises that brings to the operations world. We've not made any further progress on picking a site to say, "That's where we're going 100 percent digital." We're in the same place—we have some systems where it seems to make a lot of sense.

LaJoie: Can we reclaim analog? Yeah, yeah, yeah, we know how to do that. All-digital? We know how to do that. These really aren't tech questions. The technology is straightforward. It's overall spectrum management. There's a bunch of tools that all of us have. There's a pretty common set of issues we're all looking at. One of us may stress one path more than the other, but generally speaking, we're all looking at the same tools. The how and when of it aren't tech questions.

Davis: Let me add to that. The transition to digital doesn't necessarily have to be driven by bandwidth. If you think about having some low-cost digital set-top in front of our basic customers, which gives them access to music, VOD, the guide...the larger base of our customers are in that analog world. If we can move them into the digital world, because they want to be there, then there's a reason to transition, let alone the bandwidth matters we have out there.

There are also some interesting things we've found where we've gone with digital simulcast. One of the first things we did is, we launched the all-digital QAMs in Long Beach with 96 channels above 750 MHz. That was a mistake. We ended up moving that down to below 750 MHz to about 650 MHz, because we had MDUs out there, and wiring was a challenge. You address those one-by-one when you get to the offering of high-speed data. If you're going to cut over to an all-digital service where all the subscribers have it, we'd need to go address that proactively. It was a bit of a challenge for us and a learning issue, as well, and we won't make the same mistake in the other markets. We did it before we affected any customers. But there's a lot to be learned in this digital transition, probably more so on the operational side than on the bandwidth/spectrum side.

CED: One of the tools in the bandwidth toolbox is switched broadcast and Cox is one of a small group of operators that has tested and shared results on switched broadcast technology. Chris, based on the results of those earlier trials, what do you see as the next steps with that technology?

Bowick: I think it's a little early to be real definitive about it. I can tell you that the work that we did in Tyler, Texas (with BigBand Networks) on switched video was important work for us, and they really liked the outcome. The system would tell you, "We loved it. It worked as advertised. Very few glitches. Little difficulty in getting it launched." The bandwidth it saved was pretty doggone significant.

But there are some concerns about switched video going forward. Right now, there's one vendor. There certainly needs to be more than one vendor. Most companies don't like to put all of their eggs in one basket, so to speak. It needs to be less of a proprietary technology and more of an open technology. What we would like to do next is continue to pursue the technology from that perspective. The technology is, I believe, going to be an important one for all of us to embrace at some point in the future, because there are significant savings to be had from a bandwidth perspective.

CED: One approach being bandied around by vendors such as S-A and Xtend Networks is bandwidth expansion. Is bandwidth expansion to 1 GHz or 3 GHz one of the tools you're considering?

Fellows: We don't need to do it. Even if we did, it's not a rebuild to 1 GHz. You take the 1-Gig amps, you drop them in at 750 MHz spacing, and you eke out more bandwidth. So it's not a rebuild, it's a drop-in upgrade. I've told S-A that I'm very glad it exists. I don't need it, but if I did, I'd want it to be there. So, I think you'll see some of us—not us but some of us—doing that. But it's not going to be a big deal. This is not the start of spending another eighty or ninety billion [dollars].

Marwan Fawaz
‘We don’t think you want to make that 100 percent [digital] cutover overnight. We don’t think it’s consumer-friendly.’ –Fawaz
Fawaz: I agree. If anywhere, we'd use it in new plant extensions. The technology continues to improve, and we can take advantage of it for going-forward plant.

LaJoie: You're not going to see Time Warner expand spectrum above 750 MHz. There might be an isolated case here and there where we'd decide to do it—but by and large, no. I agree that for plant extension, it doesn't cost anything more to build the additional spectrum. So you might as well do it.

Davis: Charter is of the same mind. There are too many tools that we haven't tapped to mine the bandwidth. Charter has over 45 percent of bandwidth at 860 MHz. Whether it's advanced codecs, 256 QAM, switched digital broadcast, reduced node sizes—whatever it may be, there are tools yet to be tapped before we need to expand bandwidth.

LaJoie: There's better statmuxing, better grooming—just on and on and on. There's just so much un-mined spectrum. Bandwidth expansion really isn't on the deck. Vendors told you this? Did they tell you how many purchase orders they have?

Davis: If they can show us how to do it without digging up streets, spending an incredible amount of money, and disrupting customers, we'll listen.

Bowick: If we were to project all of our needs for bandwidth out for another five years, it's amazing how much bandwidth—given the status quo on technology and you don't do anything but just start adding services—you can come up with from a requirements perspective. I've got it mapped out in front of me for the next five years, but you could actually come up with enough bandwidth requirement that would approach 1.3 GHz. But that's without utilizing any of the advanced technologies that I was alluding to earlier. If you start looking to some of the tools available to us, for example, ensuring that our node sizes and our service groups for video-on-demand and other interactive services are less than 500 homes passed, ensuring that our multiplexing efficiency with advanced codecs is such that we can be at close to 3-to-1 for HD and close to 16-to-1 for SD, compared with where we are today, which is much less than that. Adding switched video to our repertoire for both standard-definition and high-definition, and also transitioning to DOCSIS 2.0.

I've gotta tell you, if you do all of that, there will be very little need for 870 [MHz] upgrades. Then, if you layer on top of that, the potential reclamation of some analog channels...there should be plenty of bandwidth left, and we've always said that there would be technologies that would continue to have our HFC network be the winning network.

Now, there are some costs involved in keeping our network where it needs to be and ensuring that we're using our bandwidth efficiently. But an 870 MHz upgrade, if you're offering video, voice and data services that are becoming lifeline services—and let's face it, data is becoming a lifeline service today—is not a trivial thing. It has the potential of disruption, and you have to be very careful about it.

CED: One quick conditional access question for Mike. How is your deployment of the S-A overlay going?

LaJoie: It's going fabulous and it works great. It's seamless. There's a little bit more operational complexity, but it works nicely. I'm not sure if anyone else here is doing it.

Fellows: No, but we've gone to Houston and looked at it, so I can independently confirm what Mike said.

CED: What does that mean, Mike? That you're going to do more?

LaJoie: I don't have a lot of Motorola [systems]. I'm told that we run the largest Motorola network ever built, a million devices. But other than that...

Bowick: The way I look at the S-A overlay would not be just to place Scientific-Atlanta boxes in Motorola systems. If the S-A overlay is a conducive technology to moving toward NGNA and specifically a conditional access system which is non-proprietary—including a transition to downloadable security— then that is how I would look at that technology for deployment within Cox, and not necessarily as a replacement for S-A boxes for Motorola boxes.

CED: Moving on to VOD. How many hours of stored content do you offer today, and what are you aiming for by year-end?

Fellows: We currently offer 2,500 hours, going to 5,000 hours this year. That's about 5,000 pieces of content going to about 10,000 pieces of content.

Fawaz: We're averaging about 1,500 hours, and will be at 2,500 hours by year-end.

Wayne Davis
‘If we can move [customers] to the digital world because they want to be there, then there’s a reason to transition, let alone the bandwidth matters we have out there.’ –Davis, on digital simulcast and the digital transition
Davis: We're at 1,500 hours now, and projecting 3,000 by year-end. That's market-by-market. For us, SVOD is the primary driver of hours. As we have the need, market-by-market, we may end up being over that 3,000 hours. But in general, that's where we are.

LaJoie: Our floor is 2,500 hours. Some markets have much more. It depends on which vendors they (the systems) went with. Storage is relatively easy to expand—it doesn't take more than a few weeks. Storage is not that huge of an issue.

Bowick: The number we've got on capacity right now varies depending upon how long we've been in the market, but it varies from roughly 1,000 hours to 5,500 hours in some of our more mature markets. Transitioning toward the end of the year, probably to not much more than about 5,500. Over time, that will certainly grow to, over the next several years, perhaps double capacity.

CED: How are contention rates responding to the addition of so much more free content, and the fact that customers are slowly acclimating to on-demand behaviors? Are they rising as anticipated?

Fellows: We built out at 10 percent. Sometimes we hit 10 percent at the edge. But...we rarely hit more than 3 percent. As we decouple or evolve our architecture, we can do concentration. In neighborhoods where traffic is blocked, we can add more QAMs. It's not a huge deal. It's not adding more servers.

Fawaz: We do have an advantage, because our serving areas are, on average, around 250 homes passed, per serving area. The challenge has been efficiently deploying QAMs combined with nodes. We have the reverse challenge: The efficiency of the QAMs. So we don't want to go into too small of an area as we combine. We are at six percent, but we haven't seen concurrencies go up above two percent or so.

LaJoie: It just depends on how you look at this thing, and how you build the network. We're about the same—around three percent contention. We may have seen a few higher peaks. On a system-wide basis, it tends to flatten out. Pay-per-view is great in some neighborhoods, and almost non-existent in others; the same is true with VOD. You build this out, with a number of QAM channels. Then you build service groups that pass a number of homes to achieve the desired response. Ninety-nine point five percent of the time, when somebody tries to get a stream, they get it. We've been building with content in mind since the very beginning. We keep segmenting plant. We're still combining multiple nodes on a single laser, because we haven't yet seen the contention. In some areas, we're down to one laser per node. We'll do that kind of fan-out reduction over time, as the popularity of the service increases.

CED: Does that change with HD-VOD? And how is that topic coming along?

Davis: We don't have any to offer right now. We'll do a trial by year-end, to see what the usage is. I don't see any issues with the technology.

LaJoie: The problem with HD-VOD is, there's really not a lot of great content out yet.

Fellows: In some of our systems, we offer pay movies on-demand in HD, but we're not offering any free content in HD yet.

Fawaz: We do have plans, for the end of the year, to have some HD content on VOD. But for the reasons you just heard, this is not a high priority. I think it is a significant differentiator for the industry, though.

Bowick: We have been testing it in the lab, so we are evaluating it. We do believe that it's in our future. If we look at our five-year plan, we have certainly mapped out that a significant chunk of our content is going to be high-definition VOD content, and that, by the way, is one of the other significant drivers for bandwidth in our plant.

CED: Is HD-VOD where advanced compression codecs first intersect with cable?

Fellows: For us, yes.

LaJoie: We have very specific plans for advanced codecs, but we haven't announced them publicly yet.

CED: Now's your chance.

Davis: Advanced codecs are part of our all-digital headend turn-ups.

LaJoie: You're buying boxes with advanced codecs?

Davis: No, for the headend. Encoders.

CED: Is multi-room DVR the next step in advanced features on the video side?

LaJoie: We're deploying it at Time Warner.

CED: How aggressively?

LaJoie: It's available to systems that are running the S-A guide. So right now, it's available to about 35 percent of our systems.

Fawaz: We've not deployed it yet. We have the ability to deploy on S-A, as Mike mentioned, and also on [Digeo Inc.'s] Moxi. The question we're asking our suppliers is, make sure there is full functionality in that second room—the ability to pause live TV in the second room. Also having some better storage management tools for customers, so they can move content around. Those things need to be worked out for us to go to full-scale deployment.

Fellows: Like Marwan, we've not deployed it yet. We have plans to deploy both the S-A and the Moxi boxes. Then when Motorola comes out with its platform, we'll be trialing it also.

Davis: We're testing the S-A multi-room right now, and as soon as the Moxi is available, we'll run that through trials as well. It's not only the feature set. Multi-room DVR makes the DVR home economics look much better. There's a nice driver behind that.

Bowick: It's pretty high on our list. We've waited a long time for Motorola to come around on the DVR. DVRs are just jumping off the shelf. We love the product; our customers love the product. Now we have to find a way to deploy that more ubiquitously at less cost, and that's what multi-room is really all about. That provides a huge hook for our customers. We have it in the lab today. We've identified a market, not yet announced, that we'd like to go into early this year and evaluate how that goes, and then we'll see where it goes from there.

CED: Let's talk about OCAP. What are your plans for this year?

LaJoie: We're going to put OCAP in some systems. The schedule is not firm, but at least three this year, hopefully more. By next year, we'll be rolling it out everywhere, and all the new devices we buy will be OCAP-capable. The set-tops we buy now are OCAP-capable. And for new services and boxes—all OCAP. Of course, there's the small matter of a few million boxes that aren't able to do it. They'll be pretty much feature-compatible. You'll start seeing more applications on OCAP that you may not see on legacy boxes.

CED: Is that coming out of your Mystro group?

LaJoie: Most of it. The core software is coming out of the erstwhile Mystro guys, yes. There are other companies that are building apps as well.

Bowick: We are working feverishly on both [OCAP and OnRamp]. We have been working for quite a long time, as you know, on OnRamp and expect to deploy OnRamp and interactive television in a market within the next...month to a month-and-a-half or so, and potentially another four markets yet this year. We are working very hard in the deployment of OnRamp to OCAP as well as some quick interactive services on top of that.

In parallel with that, we have been working very hard with our vendors for an OCAP stack, but also a monitor application and an extended monitor application. I've been working with our brethren MSOs to let them know that this work was also happening in the background and would love to share some of that work with them.

Fellows: For Comcast, we're a member of the OCAP LLC that is the vehicle for [Time Warner's] OCAP stack for S-A. That same venture will create an OCAP stack that will work in a Motorola environment so that we know we have at least one available. In conjunction with Cox, we have the Double C LLC, which is the vehicle through [which] we will acquire Liberate. There, we have a whole lot of experts in the OCAP world. We're anxiously awaiting the arrival of over 100 software engineers to help us launch OCAP and applications and infrastructure in this space. So we're committed to OCAP, and we're moving down that path.

Fawaz: We, too, are committed to OCAP. We're in the getting-ready stage now, bringing resources on board. We'll be doing some lab testing and some qualifications toward year-end, to be able to provide OCAP in all of our systems sometime early next year. At the same time, we're looking to collaborate with other MSOs—Dave mentioned 100 engineers— on sharing resources and doing some common testing. We do have challenges with integration and qualifications.

Davis: We're in the OCAP camp, as well. I think Marwan makes a good point of sharing resources. No sense inventing the wheel over and over again. Where we can leverage learning, we will. We're all competing for the same resources out there. We want one OCAP headend this year and scale across the footprint next year.

CED: Do you need OCAP for the July '06 removable security ban?

Fellows: No. Well, indirectly. CableCARDs need to plug into a host device. Host device specs include OCAP. That interface is under our control, so we could change it to run with SARA (Scientific-Atlanta Resident Application) or legacy software or GuideWorks (the Comcast-Gemstar TV Guide IPG joint venture). I think all of us envision a world with multiple set-top box providers. So for our own leased world, OCAP is a great way to have applications running on different hardware platforms. Secondly, there's CHILA (CableCARD Host Interface Licensing Agreement)—so we will have two-way devices showing up next year. OCAP is the way to support our CE partners. Third is, the government may force us to do it, or some other force may come in.

CED: What are your opinions on how VoIP "grows up," to be more than just a bundle component? Do you want to add on to it, with SIP-like features, or wireless?

Fellows: The short answer is yes.

LaJoie: We're already moving off of TDM (time division multiplexing) in the interconnect links and moving to a SIP (session initiation protocol) interconnect. Eventually we'll get to SIP-to-SIP. Hopefully this year you may see us do a field trial with that. I think we're all looking to move toward not only a wireless mobile application that's VoIP-based, but also seamless mobility. I think all of us are hoping that's the direction it'll go.

CED: Where are the rest of you with VoIP? Dave? Last year you talked about a specific plan. Any changes?

Fellows: We're still in the original three cities: Indianapolis, Philadelphia and Springfield, Mass. Commercially. And we're bringing other properties on line.

Fawaz: We're through technical trials and into operations trials now. We're stretching all the operational issues, from installs through provisioning, order management, and, as best we can, we're trying to get to the point where we can make it as seamless as possible—all the orders flow through; there are no swivel chair scenarios. We're testing that through trials. We're within months of launching. I think you'll hear us talk of several launches in the second half of this year. We haven't said what cities, but we're qualifying based on stringent network readiness. We want to be better than the incumbent.

Davis: We're getting more aggressive with telephony. We're very pleased with what we've seen so far, in terms of triple play. I won't speak to specific markets for competitive reasons, and we haven't given any guidance on homes passed. But it's a fair amount more than we did last year.

We're pleased with what we did last year—close to a million homes passed, marketing to a little over 600,000. We're waiting for an MDU solution, which we think is right around the corner. But 2005 looks very aggressive for voice.

LaJoie: We're launched in every one of our systems, and we've announced 225,000 subscribers (note: as of YE '04). We're moving along at an accelerated pace from there. So it's all working well, and we're very pleased with the performance of the technology.

Bowick: We're bringing five additional markets online this year. We're moving very rapidly on the VoIP front, and it's going quite well for us.

CED: Are any of you besides Mike pursuing partnerships, like Time Warner Cable did with MCI and Sprint?

Marwan: Yes. We're outsourcing our terminations and operations services to a third party [Level 3]. That helped us to market more quickly than we'd anticipated.

Fellows: We're not partnering as extensively as Time Warner and Adelphia. Obviously, when you interconnect with someone for long distance and services like that—the phone numbers are ours, the switches are ours, the third party interfaces are all part of our provisioning system. We'll also be deploying SIP-based services starting with video IM, as a service. And then over time, our SIP infrastructure and our PacketCable infrastructure will merge together. So we'll wind up with a platform that is the best of the PacketCable and the SIP worlds.

Davis: We've chosen to go it alone at this point, similar to Dave. We may do [telephony partnerships] down the road, but in the markets we're launching now, we're doing all of it ourselves.

CED: What about iTV? How aggressive will you be with that this year?

Davis: We've had an interactive service in place in a fair amount of markets, using the Wink engine (now part of OpenTV) and the Digeo-developed products. It is an interesting retention dynamic. Not anything that we charge for. As we move away from that platform and deploy the Moxi DVR, similar applications and more robust applications are on the roadmap. The specific apps that we've had out there have gained a lot of attention. Local weather, games. It's surprising how much interest there is in games.

Fawaz: For us, the games and ticker info, like weather, sports and news, are going extremely well. Very robust. We're closely working with the ETV (enhanced television) group out of CableLabs. I believe this has to be a cable-wide initiative in order for us to come up with common iTV applications. We have to come up with our own version of the red button and launch it not necessarily concurrently, but in a similar timeframe. And then finally iTV becomes much more successful.

Fellows: We've spent a lot of money acquiring and partnering on iTV. There's Gemstar and having control over GuideWorks, Liberate, and a couple more deals you'll see between now and when this is published, in April. We're fixing to do interactive applications. That's all I'll say today.