Where do we go from here? (Four CTOs bring us up-to-speed)
Heading into this year's Western Cable Show, the cable industry stands at a crossroads. It is about to embark on a historic journey that will take it into an era that includes digital video, high-speed data and potentially other services, including interactive TV and telephone-over-cable. To determine how the MSOs are preparing for this new future, we brought together four highly-respected senior engineering executives to answer a few questions related to service deployment, network construction and standards. What follows is an edited transcript of the conversation between CED Editor Roger Brown; CED Contributing Editor Leslie Ellis; Tele-Communications Inc.'s Senior VP of Engineering and Technical Operations, Tony Werner; Time Warner Cable's Chief Technical Officer, Jim Chiddix; US West Media Group's Chief Technical Officer, David Fellows; and Intermedia Partners' Chief Technical Officer, Ken Wright.
CED: What are the top two or three issues that occupy your time and thinking these days?
Fellows: In terms of hours spent, it's OpenCable, as it relates to the number-two thing I spend my time on, which is still MCNS. Finishing up the last pieces on the security model, keeping all of the vendors together so that there's some chance they will interoperate as we look at gray areas in the spec and do an engineering change request notice. Getting the suppliers to show up at CableLabs and do the interoperability testing. Preparing for this CableNet demo at the Western Show — all of that has taken a lot of time. I've spent the last two months reading responses to RFIs, putting out further questions to the vendors, reading those responses, sorting through the issues of what should be specified and how to come up with a specification that people will agree with.
Werner: I've been focused on our digital roll-out, and the practical aspects of that. Making sure the networks are able to carry the digital signals, making sure that we have headends either eliminated or interconnected down to the degree that it makes sense, to obviously avoid some costs in that area. We went into last fall shutting down the capital (spending) here, and shutting down the rebuilds. We're trying to get all the processes in place, the contracts lined up, network architectures set up on a market-by-market basis, so that we can start moving that ahead at a fairly rapid rate, starting late this year and into early next.
I guess a side point to that is that we have in the vicinity of 70,000 or 80,000 miles of plant that is HFC and close to being ready to be turned up for @Home service, so we've set up a group within my organization that's going out, visiting with the systems, and expediting the activation of two-way.
Chiddix: Digital set-top boxes, modem standards, and maybe a little light fine-tuning on fiber upgrades, I guess.
Wright: What's keeping me busy these days is budgeting. Beyond that, the main thing we're doing is rebuilding just about everywhere. Rebuilds are us! We're on the cusp of rolling out digital video in a couple of markets, and we've got cable modems out in a couple of markets. We also have a few areas where we're doing some pretty massive interconnecting of systems where we're putting in regional headends.
CED: As far as high-speed data, are you pleased with the way the MCNS process has progressed so far?
Fellows: One of the things that I think amazes all of the MSOs involved is that we get along. We actually agree on things, agree on priorities and what should be in the spec or not, and on what sort of details should be there. On the other hand, it's the first time we've done something like MCNS. And so we're also flying a little bit blind, (such as) with issues like intellectual property (and how to deal with it).
Another area is the actual compliance testing. You can either go and hook one vendor's headend to another vendor's signal, and see if it works, and if it does, you do the dance of joy in the hallway. There's (also) a view that says you've got to vary all the combinations like voltage and temperature, headend distances and number of subscribers. So we're wrestling with what compliance means. I think the industry is going to have a lot of learning to do. MCNS has gone amazingly well, but that doesn't mean that there aren't all sorts of major challenges that are appearing.
CED: Dave, what's your view on when I can go down to CompUSA and buy an MCNS-compliant modem?
Fellows: I'm sticking with the story I've had for a year. And that is, if things continue to go well, there'll be some in stores for Christmas of 1998. And if there's some unforeseen problem or issue or blow-up, then we'll miss that and it'll be the middle of '99. I still think there's a good chance that all of us on this phone call will have taken MCNS modems, we'll have proved that they actually work and there's not going to be some massive recall. Then we'll get to work on getting them into stores and figuring out how the stores make money, how the suppliers make money, how customers get signed up, how we make money, and (how to) make the buying experience easy for customers.
CED: Outside of the retail model, how long will it be until there's a critical mass of MCNS-compliant modems that are available for you guys to purchase? What's the timeframe on that?
Fellows: I think they'll show us wares at the Western Show. Some of us may test some soon after that. But maybe the security won't be quite right, or there's another revision of the chip from gate array through to custom chip that solves a few, open, remaining items. I think the real tests come in March or April of next year, and by June-ish of next year, we'll be through the trials and find that they work pretty well. We're under pressure (from) our bosses to switch as soon as possible. We keep explaining to them that a Motorola or LANcity (now Bay Networks) modem that we take delivery on today will still be working a year from now. It doesn't suddenly stop working. But there is a definite, and probably justified, sense that as soon as we can transfer purchasing, we ought to.
CED: What do you do in the interim, between now and June? Do you roll out service in a few places and then go crazy once they're all interoperable, or do you not really wait?
Chiddix: This is a good business. You keep going full speed. You roll out as business dictates. There's perfectly good product out there right now. And it's available in volume, and it doesn't become a boat anchor once the MCNS modems (debut). They can co-exist.
Werner: We are trying to expand footprint, both within markets and adding markets with existing product. It isn't so much whether Motorola's product today will be MCNS-compliant. I mean, there's no product out there today that will gracefully and truly map over. But I think Jim's point is that you can have the products co-exist on a common cable plant, at two different frequency allocations. Another strategy is that once MCNS-compliant product comes in, you migrate some of the legacy gear to other systems.
CED: What's your roll-out plan, Tony, from a timing point-of-view?
Werner: We're in the process of re-evaluating both the speed at which we do the two-way activations and our network upgrades in general. But as of right now, we're adding over 150,000 passings this month (November), and those are in the bank. And the guys tell me we'll have the same run rate for next month (December). We're hoping to actually ramp that up as we start to go into next year, taking the HFC plant that we've already built and getting it ready for two-way.
CED: Let's move into digital TV. What are your latest thoughts on roll-out plans, and how has OpenCable affected those plans, either to help or hinder?
Fellows: I'm the Luddite on the phone here, so I'll answer and get out of the way. We have plans for set-top deployments somewhere around Detroit, and we have headends installed where we think we have competition, but we are not aggressively rolling out digital set-tops across our systems at the present time. We believe that a 750 MHz upgrade with an 80- to 90-channel analog tier is compelling enough. As for waiting, it's not so much OpenCable we're waiting on, as much as prices for set-tops that we can afford, given the revenue that they bring.
Werner: We're betting the farm on digital here. And rapidly moving ahead with deployment. We're also very bullish on OpenCable, and what it has to bring, but we don't think that it necessarily affects the business today, nor is it a good reason to delay. So we've been rapidly installing the dishes and the headend equipment to facilitate availability of the signals to upwards of 90 percent of our customers by year-end. Whether or not we'll market to all those is another question. We've taken delivery of 500,000 boxes so far and have a fair number of them deployed. We've got the majority of the headends installed, but we don't have all the channel maps downloaded, and we don't have all the billing set up.
Chiddix: We're sort of in the middle of those two answers, I guess. We're proceeding with Pegasus, with Pioneer, Scientific-Atlanta and Toshiba. We expect to begin beta trials around the first of the year. As soon as that's done, they'll go to beta testing in the field, and then later be deployed in a number of our systems next year.
I think the significant thing the industry is doing with OpenCable is embracing the whole family of Internet standards, and leveraging off of all the energy of the Internet to develop and evolve open standards. Our task is really one of adapting those to the needs of the cable industry and the set-top box environment.
We view the set-top box a little differently than either Dave or Tony. It has potentially a defensive element, although there is a universe of early adopters out there who are our most demanding customers, with home theaters and so forth. Those are the folks who buy DVD machines and potentially DBS dishes. We want to make sure we don't lose those customers or the position of being their video provider.
Looking further down the road, I think digital set-tops offer us a lot more than a defensive position against DBS or merely a way to extend channels. Ultimately, cable's strength is the huge amount of two-way plant we have, with an HFC architecture. Digital set-tops open up an important area for the kind of services we offered in Orlando. And there indeed is revenue there.
Wright: Our first headend installation is in McKenzie, Tenn., which is a rural market that is channel-deprived because we've not done an upgrade there yet, and it's also a market where we have fairly extensive microwave distribution. So we decided to really challenge ourselves and have that be the first market we rolled out in. We also felt that that was a market where we were the most at risk with the DBS Christmas buying season. Beyond that, there will be more systems that will be firing up between now and the end of January, (including) Greenville/Spartanburg, N.C., followed closely by Nashville.
CED: What does your capital budget look like for next year? Will it be more or less than 1997?
Chiddix: For us, it's pretty much even keel. Our overall capital spending is about $1.6 billion next year, which is very similar to this year. That basic level of spending will last to the year 2000, when we complete the last of our upgrades. That capital is broken into upgrades, set-tops, maintenance capital and new business, which is cable modems and such. After the upgrades are completed, we expect our spending to fall to under $1 billion.
Fellows: We're spending about $1.5 billion this year, and (about the same amount next year). You know, it's flat, but still we're spending a huge amount of money on upgrades.
Wright: For us, '98 looks pretty similar to '97. In late '96, we set out to do what for us is a pretty widespread rebuild program, and '97 and '98 were the two major years. It's nearly $350 million for '97 and '98 combined, and we'll have spent about half that by the end of this year.
Werner: We're still sticking with our plan of record. As these joint ventures we've announced get closed, they'll slide out of our plan. At the same time, we're looking to accelerate our schedule for the remaining properties. My guess is that we'll shift a lot of the 2001 and 2002 spending into the '98 to 2000 years. So our numbers right now are about $760 and $850 million for next year, and between $950 million and $1.2 billion for total capital — that includes maintenance and line extension — and upgrades may grow as we move into the upcoming two years.
CED: What percentage of your plant is currently two-way HFC, and how much will that grow next year?
Wright: At this point, we're at just under 50 percent of our plant that's two-way HFC. And by the end of next year, we'll be at the 90 percent mark.
Fellows: We are just over 50 percent where the bandwidth has been upgraded to either 750 MHz or 550 MHz, and my calculator says 26 percent are two-way HFC, data-ready homes-passed. That will grow 20 percent (per year) for forward plant bandwidth, and we'll actually end this year at about 60 percent. (On the reverse plant), we'll end the year at roughly 30 percent, and that catches up at a little bit more than 20 percent increments going forward.
Chiddix: We're at a little over 50 percent as of the end of this year, that's both two-way active and HFC. Right now, we also have non-HFC plant which is two-way active. We're mostly tracking the HFC upgrades, and currently we have 18 million homes passed. A little over 9 million homes passed that will be upgraded this year, and that grows at a rate of 3 million homes per year until we're completed at the end of the year 2000.
CED: What about the return path? We keep hearing how difficult it can be to operate there. Is everybody planning to use filters for the upstream path, or going the route of clean plant?
Chiddix: We're in favor of clean plant. We've used filters in some areas where we needed to clean something up that wasn't quite up to snuff. But we're firm believers that it is quite possible and not horrendously difficult, in fact, to clean up plant to the point where you don't need to put filters everywhere in order to fire up two-way services. Adding filters is a crutch that keeps you from running plant the way that you should run it, and they become more and more operationally complex as you go down the road and get higher penetration of two-way services.
Wright: I'll just say ditto to that, Jim. We're really in the same boat. Across the board, we're firing up two-way plant without filters, and finding that we can make it work. As Jim said, if you use filters, you become your own worst enemy, because penetration grows, the filters come off, and problems then start showing up after you bring customers on.
Werner: We're leaving those decisions up to the systems. They can make that call. We obviously work with them and advise them, and for that matter, we consider filters to be a useful tool. CableLabs has done some interesting work that clearly proves that filters are a valid tool. They showed that filtering can and does remove greater than 90 percent of the interfering types of noise and burst energy. So we think filters are a good tool, but we also have plant that's very clean and doesn't need them. But we're not religious about it one way or another.
Fellows: Dave "Band Aid" Fellows here uses filters in five of our six regions. When someone signs up for a two-way service, the filter actually gets moved to one side of a new splitter we put on the house. In the sixth region, we use reverse equalization, if you will. We add loss selectively in the reverse plant, so that the transmit levels from all homes are roughly the same and roughly the maximum allowable.
Wright: We're doing that, too, in our design. We're padding at the node to bring our worst-case terminal devices near the top of the output range.
Chiddix: Yes. Reverse equalization is a very clever solution, and we're experimenting with taps that have built-in return equalization.
CED: It seems that RF telephony, though, is seeing some momentum, at least if Cox is any indication. And Tony, you say you continued to be impressed with TCI's telephony experiences in Hartford.
Werner: We don't have any further plans for it right now. We'll stick with the experiments we have in place.
Fellows: On RF telephony, we're in the process of selling our interest to Cable and Wireless, but we have over 20,000 paying telephony customers in Australia. In this country, we have some trials going on in Atlanta, but have not made the decision to roll it out yet. And then back in my role, I'm placing bets. I'm making sure that I'm not so committed to HFC telephony that if IP telephony does what we hope it does, I've not done anything stupid. On the other hand, if IP telephony only turns out to be an Internet ham radio equivalent and does not have the quality or functionality to send faxes or receive voice mail or DTMF tones, then I haven't not entered the phone business waiting for this Holy Grail. So I'm trying to balance the effort and the investment here.
Chiddix: We have thousands of HFC telephony customers in Rochester, and are very confident that the technology works well and is quite manageable by regular cable folks. But we are not deploying it elsewhere, because we're quite concerned with the regulatory structure around competitive residential POTS. It's not a very good business right now.
Wright: We're in a position where we have to watch very carefully what we do with our capital, and make sure we make the best use of it. So we made the strategic decision some time ago not to put the extra capital that's necessary into our plant to start doing HFC telephony.
CED: What about IP telephony? Does that interest you?
Werner: We're bullish on that, also recognizing that there's a few things between us and the delivery of that service. Standards for the networks, and working out the terminal equipment. But I'll echo what Jim said, which is that we're finding on the HFC side that our networks can do it, and that the technology works pretty well. Whether or not it's in the long-term cards for the company is another question.
Chiddix: One of the things that's intriguing about it is that you may be able to leverage the same infrastructure that you build for the cable modem business. But you can do teen lines and second lines and so forth with the HFC gear that's available today, as well as IP telephony. IP telephony does have another capability, which is that some of the expense can get built into the PC, thereby reducing costs, but that also very much colors how it's used.
Fellows: One of the problems is that IP telephony means a whole bunch of different things. It could be viewed as an alternate way to get ones and zeros out of the home. And to that extent, you do worry about lifeline telephony. But we are powering our cable plant so that in times of power outages, the plant still works. With IP telephony, you have to worry about a box in the home; however, I have in the lab here at the Pilot House a gateway that has an RJ-11 jack and an Ethernet jack that doesn't consume very much power. So you can imagine them lasting as long as a cordless phone or surviving a couple-hour outage.
IP telephony is also a way to get a whole bunch of features that you can't get with a 5ESS switch and a normal telephone network. Like, you're looking at a Web page and you click on something, an icon, and a phone call is connected. Or, (when) you get an e-mail from someone, you hit a button, and you've got a telephone connection back to them. If they don't happen to be there, then you e-mail a voice mail back to them. And there are other convergence kinds of features.