Probing the minds of cable's braintrust
Five of the industry’s best and brightest discuss their agendas for advanced services such as iTV, VOD, HDTV, VoIP and home networking
Now, perhaps more than ever in the cable TV industry's storied past, the future is unpredictable. Competition is stronger than ever. Economic pressures are extreme. The industry has spent billions over the past five years to upgrade its networks, making them bi-directional and more reliable than ever. But beyond that, the industry is striving for consensus. What new services will ride over those rebuilt networks? Which ones have the best chance to provide new revenue? Is IP telephony finally ready for prime time? To provide some insight into the latest thinking in cable company boardrooms, we caught up with the industry's technical braintrust and put them on tape. Included in this virtual roundtable is Comcast Corp. Senior Vice President of Strategic Planning Mark Coblitz; Cox Communications CTO Chris Bowick; Insight Communications CTO Charlie Dietz; and Senior Vice President of Network Engineering Paul Gemme, and Senior Vice President of Subscriber Technology and Advanced Engineering Mike Hayashi of Time Warner Cable. What follows is an edited and aggregated transcript of those conversations.
CED: Looking across the voice, video and data landscape, what are your major priorities in the short-term?
Other things that will flow on top of that would include things like tiering. (Exploring whether or not) there are some tiered services that we would like to offer our subscribers, which we could not previously offer under @Home. The question of open access is out there looming, and we've had a trial ongoing in El Dorado, Arkansas with Earthlink and AOL just to insure that the technology is ready and scalable.
On the telephone side, we want to add some advanced call functionality this year (to our) circuit switched (product). Things like talking call waiting and talking caller ID. We've also got an initiative relative to voice-over-IP. We're not keeping our heads in the sand relative to voice-over-IP, (so) we'll be doing some trialing this year.
In Oklahoma City, we have placed a GR-303 gateway between our existing circuit-switch and potential voice-over-IP customers. That allows us to leverage that switch, and provides VoIP infrastructure on the line side of that switch, and perhaps can allow us to expand the geographic reach of that switch.
On the video front, the primary initiative is video-on-demand. We previously announced that we would be in seven markets, to about 43 percent of our homes passed by the end of the third quarter. So that's a major initiative for us to get those systems launched. We also want to be more aggressive on high definition television.
We've got HDTV already up and running in Omaha, and we've got it running in the lab and it's about to be launched in North Carolina.
Another, of course, is interactive television, and we are working with Liberate as a middleware vendor, and MetaTV on what we would call an iTV Lite, broadcast-only version of interactive television to offer news, weather, and sports.
We'll have some trials relative to iTV. The only one that we've actually talked about in the past has been San Diego. That has been with Liberate on the S-A (Scientific-Atlanta) platform and we have been working with both companies to work out the kinks and the bugs relative to a launch of that kind of service on that platform.
We're also very interested in PVR (personal video recording). We are bullish on the technology and the need to have an integrated PVR solution for our subscriber base. We're working with (Scientific-Atlanta and Motorola Broadband Communications Sector) to ensure that.
We're (also) working on the solidification and expansion of our high-speed data networking, which we've just transitioned from @Home. We're working on future set-top designs, and we're rolling out high definition television in a lot of our markets. (Looking) out a little bit farther, (we're also interested in) home networking. Not the networking of multiple PCs, because that's an operational thing that we're clearly working on, but what might be called a media center-type application–how we move information around the home and extrapolate beyond the kinds of technologies that we've developed in DOCSIS and PacketCable and OpenCable, and find ways to make that more robust, and add more value for our customers in the home.
CED: Let's talk about DOCSIS 1.1. To what extent are you all transitioning to that technology?
Coblitz: We have a small lab where we can test products, and we've done that with a whole series of DOCSIS 1.1 equipment. It's certainly our intention to be buying DOCSIS 1.1 equipment. Today in my lab, we have a large number of call agents for IP telephony.
It's pretty likely that we will start acquiring 1.1 CMTSs, put them in particular marketplace(s), move CMTSs that happen to be there to other places that need more capacity, and then learn all the operational sets of issues that (come) with the increased security and privacy, QoS, and the way we handle monitoring the network. And then once that's under our belt, continue to move 1.1 into other marketplaces.
Gemme: We're moving forward to DOCSIS 1.1. We're evaluating some of the equipment now. Now we have to do regression testing on the software back in Herndon (Va.) to make sure that it's compatible with everything we have in the field, but we feel that we're going to be able to implement that pretty soon. And of course, weaving into the voice-over-IP category, 1.1 is essential to launch (IP telephony) in any significant fashion.
I guarantee that by this summer we'll be doing trials on at least one or two CMTSs that are 1.1-capable (or) qualified. The goal is to get 1.1 rolled out (by the end of this year) because, if you're going to move into Voice-over-IP, at some point, you're going to need that in place.
CED: Once DOCSIS 1.1 gets deployed, and guaranteed quality of service is available on the network, what kind of tiered service do you anticipate investigating?
Bowick: We would certainly be looking at a tiered level of service to provide more choice. The question is, what should that choice be? And I don't know that we know the answer to that. In our discussions internally we have not yet finalized the price, or an actual tier structure, but what we want to do instead is some controlled experimentation on price points and tier levels, so that we understand what it is that our customers want and need.
Dietz: Well, there's definitely the small office. With @Home controlling our networks in the past, it was very hard to get into that business. They controlled all the IP addresses, and the routing, and we didn't have access to our own clients. Our new arrangements (are) designed for multiple ISPs, so we'll be able to open up to multiple ISPs. It would work the same way for a large enterprise business in our market. If we have a company that has a strong presence in our market and wanted their employees to work at home and be able to route back to that enterprise, the partnerships we've got today will allow us to do that. Same as we would route traffic to another ISP, we could route directly back to that business presence. That's going to give us great utilization of our bandwidth during the day. With 1.1, we'd be able to set up a level of service for that customer during the day, and still give us the ability at night to shift that traffic when the kids get home from school, so they could do their Internet surfing with a different level of service.
Coblitz: We're beginning to experiment with different forms of streaming media that would be (sent with) QoS over 1.1. There's a belief that (there's) some amount of video that's going to be occurring over that IP structure. So the question really becomes, what are the sets of issues and dealing with this over that IP structure, and that's DOCSIS 1.1. I don't think we're at a point yet to talk about revenue models and cost models, but rather thinking about how this stuff works, how we understand all of the things about delivering streaming IP over our networks. That will extrapolate into the requirements that we're going to have to place in order to (do) that all the way through to other devices within the house.
CED: What's going on with telephony? What issues are you grappling with and are you testing IP-based telephony?
CableLabs has no 1.0-plus certification or qualification of product at all. Although we are doing a couple of trials (with) 1.0-plus software applied to the CMTSs, that was something we did to get some quality of service. Because you are talking (about) a lot of small bits when you are doing phone traffic, and you don't want that to get interfered with or overrun by high-speed data, you have to set aside specific bandwidth for those calls. And that's what 1.0-plus did. It handled that a little bit better than 1.0 did. With 1.0, you didn't have any type of quality of service, so it was (random) as to whoever got on, and the modems would slow down. But now we are able to throttle modems and everything, and assign bandwidth, so that's good.
Dietz: Right now, I think it's a little early. We need to better define and roll out DOCSIS 1.1 and PacketCable. We looked at it a year-and-a-half ago and said the technology is not ready, so we chose to go with AT&T and go with the constant bit rate product. That said, I'd love to get to IP and have that efficiency so that I'm not carving bandwidth out for every individual application, and some cases having it sit there unused.
We've got three (constant bit rate) markets up now, and a fourth one sitting waiting for regulatory approval. They all (use) the Arris product. We will be building off all of those markets and keep expanding off of them. The next new market we go into we'll definitely look (to) see if IP is ready at that point, and then if it is, go with it.
We're going to stick with the primary-line model. We are putting in standby powering, and centralized powering, so everything is backed up to a gas generator to get that 911 reliability.
CED: Charlie, do you have any plans at this point then to even do any lab or field trials with IP telephony product and service?
Dietz: A lot of it is going to depend on AT&T, too. Our agreement with them is that we're responsible from the headend to the customer, and they are responsible from there and back out to the switch. I'm sure they are going to definitely want to be involved because a lot of the cost savings and efficiencies are going to be on that switch side. We are fortunate to have them as a partner. They are doing a lot of development and testing of those products, and when they're ready, they'll be coming to us to help pick a market to roll them out in.
Gemme: The approach that we anticipate is a second-line service. There's a whole plethora of reasons for that. First of all, the ILECs will just shoot right back if you try to do it as primary-line service. We don't want to put in power backup. Power backup is very, very expensive. There are a couple of MSOs that chose to go that route, and I don't know how they are coming off those costs. It's not only backing up the plant, but it's also backing up the powering to the homes. So you have to power to the home from the plant, and that's where it gets real expensive. If your goal is a 30 percent penetration, you need to back up your plant to be able to handle 50 percent or 60 percent because, just like any other service, there's going to be lumpy distribution.
We're not going to power the unit in the home. They have an option to do a battery back up for the unit in the home. (We're) also looking at a "magic" phone that will not only be a phone in the home, but as you go outside and go past the reach of the device, maybe it switches over to cell. It's a possibility. We all know that cell phones are pretty reliable at this point, and as a secondary form of reaching 911, it's not bad. When I say we have a secondary-line offering, we will have 911 capability. You will be able to dial 911 with that phone. The problem is that if the power is off in your home, then you may not. But with the magic phone, then all of a sudden, you'd have power to that phone.
But the truth be known, you'd be surprised how many homes don't have primary-line service. Because they've all gone to phones that require AC from the house to power them. When your power goes out, that phone is out. The phone companies have not been forthcoming in telling their customers this. So there are a lot of people sitting out there fat, dumb, and happy, and they don't have 911 service if their power goes out.
Coblitz: We just finished testing IP telephony in Detroit. What's being tested is IP from the headend to the home. The IP that is being tested there is not (with) call management servers, which I have in our laboratory here. (In Detroit, we have a switch) because that was a MediaOne property, and it came with the 5E, and you want to look at how you can get some savings by using IP from the headend to the home, and there are some significant cost savings in that part of the equation.
(From that test, we learned) that DOCSIS 1.1 is carrier-class. We proved that this technology can technically work, that the cost structure is several hundred dollars less in capital, but to be workable, that there were alternative ways of powering other than powering the network. And I think the quality that was delivered to the customer was not much different from the circuit-switched (product). It's a very positive outcome.
CED: Turning to set-tops, what is your strategy for rolling out digital boxes and other, more functional, boxes to consumers?
When we look at technologies like Moxi (now part of Digeo Inc.) and video- and audio-based home networking, I think it's very interesting. It's probably not something ready for this year, but it's certainly around the corner. We started the (Explorer) 8000 effort (approximately) two years and four months ago. I would say that's really unacceptable going forward. We need to have products that could come to market inside a year. Having said that, we're excited about the 8000. Gemini (Time Warner's internal name for its PVR-capable Explorer 8000 set-top) is in full development. We are going into field beta testing soon, in the next 30 days or so. You heard it here first.
We should be aggressively pursuing HDTV set-tops. When you walk into a retail store today, and if you ask for the premier choice content for HD, nine out of 10 times, it will be satellite. It's really ironic, because we have more bandwidth. So we should be the preferred choice for HDTV content. So, somewhere along the line in the last 12 months something happened. I'm just being a little cynical about it, but we should be deploying HDTV content. There's plenty of content. And we have the bandwidth advantage. It's not that expensive, if you think about it.
We have digital deployed in every market. But we would like to hit 50,000 HD customers, which is a very conservative number we think, by the end of this year–hopefully double that.
Dietz: We've got the Motorola platform, the standard (Motorola) DCT-2000, and that's going to be our workhorse for a long time. On that box today we've got a server-based interactive guide, we've got our local switched product, which has a lot of local content for each individual market–school menus, restaurant menus, what's at the local theater, applications like that. We've also got a string of very simple games that customers seem to love.
The majority of our markets now have video-on-demand enabled. We rolled (out) our first systems at the end of 1999. We've got almost 25,000 streams out today. Video-on-demand comes along with our gateway digital product. So it's there for the customers to use. With the same gateway product you get access to your full multiplexes of HBO and Showtime and services like that. People seem to like the product; it's being accepted.
I do see the need for a more advanced box. We'll be getting into some HDTV by the end of this year. And then PVRs integrated right into the box down the road.
Liberate (the middleware supplier) has been very good to work with. Where it gets to be a bear is all the testing between different phases of boxes. As the manufacturer finds different ways to keep costs down inside the box, we've got to constantly check the software to make sure that it integrates correctly with every new phase of the box. The last year-and-a-half or so, we've seen a lot of different upgrades of the box. The good news is, that has stabilized. We'll see one new phase of DCT (set-tops) this year that we'll have to integrate to. That said, Motorola has been super. They've got their Acadia test lab to give us a warm and fuzzy feeling about the new hardware and software we give our customers.
As for HDTV, we will charge for it. The box is going to be more expensive to us. Initially it will be the HBOs, the Showtimes, or (other networks) that have cable programming up there. We are starting to talk to the local broadcasters about carrying their HDTV service in our market as well.
CED: Speaking of video-on-demand, what's everyone's agenda related to that? Charlie, what can you tell us about the take rates you've been seeing on video-on-demand? Is it more or less than what you thought? Are you going to have to scale up faster?
Dietz: I guess it's a two-pronged story. When we launched several years ago (customers) had buy rates in the 200 percent to 300 percent range, which meant every digital box we had out there was buying two and three movies a month. In the last year or so, it's fallen off, and part of that has been (attributed to) the studios pulling back the product. We're hoping that that's going to open up and we'll have the product that people will want to buy. Those are deals that have to be worked out with the studios.
I've heard my boss say that the buy rates we've got are a little over 100 percent per month. However, when you look at it, it's 15 percent of the customers actually using the VOD service, which equates to those customers buying five or six movies a month. So I think that the people who get into it and start using it really love it.
We are going to do other things than VOD and those types of services–maybe do something with a sports provider (for video) clips. Something that will get the customer to go to that at a lower entry point, and get used to that functionality, because once you get used to it, you won't go back.
Bowick: In Hampton Roads (Virginia), we've got it up and running to a couple thousand subscribers. We also have it running in San Diego on the Scientific-Atlanta platform. Hampton Roads is on the Motorola platform. And we're using Concurrent (servers) in both of those locations. I don't like to be single sourced, so we are continually looking at our VOD vendor base, and looking at potential second sources. I will continue to look at them and see if it makes sense to make that move.
We've been focusing on deployment of the technology and the software associated with that technology, and ensuring the scalability of the platform before we start rolling it out more broadly. This is the year we'll roll it out. At the end of this year, we're not going to be much different, from a footprint perspective, than any of our counterparts. I'm feeling very good about where we are (with) VOD right now.
Because of the way we are set up–with a master telecommunications center and several hub locations–in most of our major markets, the distributed architecture seems to be an appropriate architecture. Now you can envision, over time, a migration of heavily-used titles toward the edge of the network, and library titles toward the center of the network. So a distributed architecture certainly works in that regard. We have been working with N2 Broadband on the pitcher and catcher's mitt scenario for distribution of content via satellite in the background to the distributed server farms out there.
CED: Chris, at one time you were trying to make VOD work on some DCT-1000s in Hampton Roads. Is that still alive and well, and how is that going?
Bowick: We do have DCT-1000s out there. They are 64 QAM boxes only. Our preference for VOD, for spectral efficiency, is to launch with 256 QAM, just as it is with high definition television. We want to launch with 256 QAM just to improve the spectral efficiency. So that is certainly an issue, and we're talking to Motorola about what to do about the 1000s. But that's about all I can tell you.
Gemme: We also envision the rollout of video-on-demand as being a migratory thing. (Maybe) you begin with a centralized server in a smaller division, and understand that eventually you are going to find a hub site that has more business than the others that might overload your central server. So maybe you move a server out there to take care of that. The centralized server then maintains the other three hub sites or whatever. In larger divisions, you're faced with the task right up front, but you've got to deploy a lot more set-tops, (because) you're going to have a lot more events occurring, etc. It depends on the division. And just like we said with everything else, you are not going to sell these services ubiquitously across the networks. So as you begin to run out of streams, you (ask), where's the next logical place to put another server? How many streams will we need to support them? Everybody asks, do you centralize or distribute? I think you do some of both, initially. It depends on the size of the network.
Hayashi: And don't forget about session creation and breakdown, which is a related issue to traffic management itself. Movies are interesting, because you charge for that movie, and when you buy that movie you are committed to watching it. But with HBO on Demand, we observed behavior where (subscribers) went in and out by a factor of about 10 times. They were in and out on the average of about eight minutes or so. One of the big issues is having a much more scalable session management environment. We just went through a very big upgrade with Scientific-Atlanta to what they call the System Release 1.5. That, so far, has scaled substantially beyond the Columbia (South Carolina) experience (which overloaded last year). So, from a session standpoint, I think we're going to be okay.
CED: Chris, have you done subscription VOD in any of your VOD tests yet?
Bowick: No, but I can tell you that within the 43 percent of our homes passed (by VOD) by the end of the year, several of the markets will be doing SVOD. (We'll offer SVOD) in a mix of large- and medium-sized markets.
CED: That's interesting, because some people tell me that VOD doesn't really scale for big, metro systems. You haven't done SVOD, so you haven't seen the usage statistics. But do you have any concerns that VOD is not going to scale? That you are going to run into problems?
Bowick: The reason we've done it the way we've done it is so that we can work through all of the issues that you've talked about without impacting customers–so that we can understand the scalability issues. That's why I said we have not launched very broad and very thin across markets. We opted to focus in two markets, work the details out, understand the network implications, understand the scalability, and when we're ready to rock and roll, we rock and roll. And that's what we're doing this year, and I think that you will see that having done that homework will pay off for us this year as we launch in those other markets.
CED: What about personal video recorders? There are a variety of ways to deploy them–either through digital set-tops, via a separate sidecar or via a centralized, network-based server. Which way are all of you leaning?
Bowick: I mentioned PVR earlier, and we certainly think that an embedded PVR is something that we would really like to do. When you look at the Moxis of the world, or the next-generation Replay, or SonicBlue technologies, or the next-generation TiVo, which are really home network asset management systems, those are all very intriguing to us. I think the concept is certainly a good one. We have looked at Moxi. We have looked at the others as well, but we have no trials planned, although we do have some work ongoing on the PVR front. But I look at Moxi as the next generation of PVR, which is complete electronic home asset management. We're looking at all of that.
Hayashi: We certainly would like to see "Maestro," our network PVR, happen, from the perspective of offering your TV anytime when you want it, without investing more in new boxes. The network PVR solution that Jim Chiddix (President of AOL Time Warner's Personal Interactive Video Group) is working on is something we look toward. But in the interim, I think we need to deploy these PVR boxes. We specified these boxes two-and-a-half years ago, so we are anxious to get them out.
Coblitz: The question for us is what's the timing. We did a trial with TiVo some time ago, in which we learned that people wanted integrated devices. They like it, but they want it integrated.
I'm not suggesting that's what we're doing. I'm just saying that's different because, although it may be two physical pieces of hardware, they are fully integrated (with) one another. I would say that the second thing that is not yet clear is the business model for the delivery of this. We know the satellite guys have been putting it out, so there's a business model that's there, and that's something that we're working on as well.
There are a lot of advantages to network-based PVR, not the least of which is the potential to deliver services to an awful lot of homes that already have digital set-tops in them. There's got to be an awful lot of technological development to make that happen. I'm also a big fan of IP structures as a way to deliver this, which, of course, existing set-tops don't have. So that's something which has to be worked out.
That being said, we know there are some subscribers who will want to have control in their own home, and there's an interesting advantage for us if there is storage in the home. I think we're going to find a mix of capabilities, starting with VOD–which is an interesting and very important first step in having consumers get what they want when they want it–and then extrapolating that concept into delivering more content over our network, and balance that with people who decide that they want to have storage in their homes for more personal control.
Hayashi: Paul (Gemme) and I have been doing this for quite a while, and we have this bias. We just have not done well with sidecars in general. There are a lot of different theories as to why it's hard to market a sidecar. One is that you have to go back into the same household and undo their wiring and so forth, and it's confusing. I think that's a big issue. But we understand the great sense of urgency to do something right now. So the temptation is to just slap on a sidecar, and put on a little IR blaster, and it should be okay. Could be, (but) we're not sure. We're not pursuing that route.
CED: Can each of you tell me a little bit about interactive TV and your plans for that application? Charlie, what have been the real winners? Are there any that seem to be really hitting with the customers?
Dietz: VOD definitely. The customers who embrace it love it. The interactive guide, while people don't realize it's interactive, (allows customers) to be able to go in, punch through things and get local information–they view that as a big plus. You never know what's going to pique a particular customer's interest. What one customer loves, another will hate, right down to the clock on the box. One customer will say, "Where did my channel number go? I've got this clock there now." Another one will say, "Wow, this clock on the box is great! How come I didn't have it before?"
Bowick: We certainly are looking at Internet to the television as one potential application in the future, but our focus, at least this year, will not be that. It will be more of a focus on news, sports and informational kinds of services that our customers may want overlaid on top of the television set, but not necessarily launch out to the broadband Internet.
CED: Chris, do you have any plans to do any commerce, walled garden, stuff like that?
Bowick: That's not a large focus for us this year, although you need to understand, a couple of years ago those were very big buzzes in the industry. I mean the Internet was a craze in the industry, and the thinking was that there are a lot of dollars around those areas. But that may not necessarily be true today. So we're trying to focus on things that our subscribers may want from an informational and content perspective, as opposed to e-commerce early on, and then e-commerce would come at a later point when you get closer to the Internet-to-the-television model.
CED: Are there any services you'd want to drop because you just didn't find any penetration at all with the customer?
Dietz: No. Even with simple games, we've found customers are in there using them. There will always be different screens on different applications where we're going to have to look to see what customers are using and decide whether to keep things like that on, but, has there been one single thing that customers just hated? No.
CED: Mark, what about things like online gaming? I know you have a young son who's excited about gaming. How does that translate into things you'd like to do over your networks?
Coblitz: I have a Play Station 2 and an Xbox (from Microsoft). I have multiple digital set-tops in the home because we like all that content in all the places that we work. It has TiVo, and I'm about to install a Replay so I can play with that as well. And then we have a high-speed cable modem that's 802.11b linked to my son's PC.
And my wife has a laptop. The PCs that both my son and I own have all of the video stuff so that we can experiment. So we can experiment with everything from writing our own DVDs, to taking other content off the 'Net.
And the high-definition television set that I have has component input because we're not into the digital world yet. So I have component analog, which is great and delivers wonderful quality product. But the rest of what I get from our digital set-top is an unintegrated solution that has to be seen through the regular analog input. I have to switch back and forth.
I can tell you that as a consumer of all that, I really want to see it all in one place. It's quite frustrating, actually, to be watching something in high-def, and to want to get guide data, and not be able to get it. It's those kinds of things that I believe you have to learn to do on your own. And frankly, one of the reasons that I got the Xbox was because of how great it is to see in high-definition. It's very incredible in high definition. It truly is. They've done a good job.
CED: Is your Xbox hooked up to a cable modem connection?
Coblitz: It's not yet because they haven't done that yet. But I think you can expect that somewhere over time I'll try that. Or my son will try that. Actually both of us will try that.
CED: One final topic: home networks. What's the latest thought there? Can cable operators make money in this area, and if so, how?
Bowick: We are looking at home networking as a potential opportunity. As you know, our subscriber base is doing it today on their own, using Linksys equipment. I've got one in my home. The question is, can we make a business out of that? So we are, in fact, working with a couple of companies to try and figure this out. Can we provide this as a service to our subscribers? Can we say, "Look, Cox is your in-home service provider. We will set up your home network, maintain it for a fee, we'll take care of any and all trouble and service call issues." Is it something that we can and want to manage as a business? We are launching a trial to determine that. We'll trial a price point here or there, and see if it's a business for us.