After falling off the face of the earth about three years ago, the concept of fiber-to-the-home (FTTH) is again gaining strength. For instance, Scientific-Atlanta Inc. recently invested $8 million in Alloptic Inc., a company that is looking to expedite FTTH development. According to some FTTH vendors, the cost of driving fiber into the home is dropping so fast that greenfield operators and incumbent local-exchange carriers may be able to implement it-without going broke, that is. Meanwhile, as interest in video-on-demand and video streaming continues to rise, just about everyone else agrees that the thirst for bandwidth has never been greater. So, where does that leave cable in the FTTH equation? Is FTTH affordable, or is it still pure fantasy? Will cable operators even be required to deploy FTTH architectures in the foreseeable future?
broadband editor Jeff Baumgartner discussed those issues with Alex Best, executive vice president of engineering at Cox Communications Inc., an MSO that plans to finish across-the-board network upgrades by 2002. The interview was done prior to Best's announcement that he would retire at year-end. An edited transcript follows:
MCN: To begin, what's your view on fiber-to-the-home and the implication, if any, it could have for the cable industry in the near-term?
Fiber-to-the-home vendors don't call on cable operators. They know we have a piece of coax going in the home. Therefore, they haven't felt compelled to solve a last-mile problem.
Fiber-to-the-home is really the Holy Grail. As an engineer, it certainly sounds neat. Fiber-to-the-home is the end-all to the bandwidth equation, but no one's been able to cost-justify it.
Even if you could do it, why would you do it? It's kind of like building a six-lane expressway up to your driveway. It's nice, but it's not necessary. Since we have a hybrid-fiber coax, we still maintain that it's the most cost-effective and, in fact, offers enough bandwidth to do everything that we can imagine going forward.
Whenever someone asks me about fiber-to-the-home, I always ask, "Is that a home-run fiber-to-the-home architecture?" meaning that I have dedicated fibers from my headend to your house, or, "Is that a tapped-fiber system?" Usually, when I ask that question, a whole lot of discussion happens, and I never get the answer to that question. Fiber-to-the-home means you'd better have a lot of hub facilities.
MCN: Does adding those hub facilities only exemplify how expensive fiber-to-the-home would be for cable operators to deploy?
Sure, absolutely. Those cost money, because they're bricks-and-mortar. But if you don't have lots of them, then a home-run fiber system is out of the question. Our fiber cable counts get pretty hefty coming out of our headends, which take fiber down to 500- or 1000-home nodes. We have fiber cables leaving the headends with 200 or 300 fibers, so you can imagine how hectic it would be if they were home-run to the home. I don't think anyone's advocating home-run fiber to the home.
MCN: We've been hearing a lot lately about fiber-to-the-home from the equipment vendors. Obviously, they're biased and they want to sell those products. Some say fiber-to-the-home is ready for prime time now; others say it will be three to five years before it becomes cost-effective enough to implement. With that to consider, do you think fiber-to-the-home is a possibility for the cable industry in that timeframe?
Personally, I don't. Like I say, they're not calling on cable companies. I'm sure they're calling on Bell companies. I do keep an eye on fiber-to-the-curb architecture, such as what went on with U S West's VDSL trial in Phoenix using NextLevel Communications. That's fiber-to-the-curb, serving in the neighborhood of 25 to 30 homes. In the end, they basically said, we're not going to do that anymore. It's too expensive.
I visited the VDSL trial in Phoenix, and their fiber terminated into a rather large pedestal with very sophisticated demultiplexing electronics inside in order to get the multiple-hundred megabit video signals and separate them out into the individual video signals that went into the home. That's a costly proposition.
Our fiber is shared by 600 or 700 homes. We think ours should be justified by sharing the fiber costs with that many homes. (U S West's) trial tried to share the fiber with 25 to 30 homes, but it couldn't make the economics work. Therefore, how can I assume that fiber-to-the-home will work?
And it's not only just the cost of the fiber. Let's face it, the appliances in the home still have F-connectors on them. When you transition fiber back to electrical, that means you'll need a device. In our case, that device is the node, and a node costs us $4,000 on average, but it's shared by 600 or 700 homes.
In addition to the glass, the challenges of fiber-to-the-home also include the transition of the light energy back to electrical energy, because all of the appliances in the home are expecting electrical energy.
I remember 12 or 14 years ago I spoke to a group of Bell company people and they asked me, "Alex, when are you going to take fiber-to-the-home?"
My answer was, "If you go to Circuit City and buy a TV set that's got a connector in the back that says, 'Screw the fiber in here,' then I'll take fiber-to-the-home." My point was that the cost of transitioning from glass to electrical is still an issue. Today, we share those costs.
Fiber-to-the-home vendors don't call on cable operators because they know we already have a broadband pipe into the home, and we'd be the last people to push fiber all the way to the home.
Bell companies, which have a pair of copper wires that basically doesn't support what they want to do, would be the most likely candidates for fiber-to-the-home.
MCN: So, at least for the cable industry, does there appear to be more hype than reality when it comes to fiber-to-the-home?
If someone came to me and said, "Here's a great fiber-coax architecture, but lends itself to the next step to take fiber-to-the-home, and here's how you would do it and here's how much it's going to cost you," I'd love to hear that conversation. But they just aren't pursuing us. They're pursuing companies that have an absolute bottleneck in the last mile. And, of course, that's the Bell companies.
A dead giveaway would be the so-called overbuilders, who are advocating competing with cable operators. I've looked at some of their business plans, and I haven't seen a single one of them advocating fiber-to-the-home. Everyone of them is advocating hybrid fiber-coax. Here's a group of companies starting from scratch, and every single one of them is talking hybrid fiber-coax. There's got to be a clue there.
MCN: Still, aren't those nodes typically serving a very small cluster of homes?
MCN: One hundred or less.
I've seen some advocating that, but serving 100 homes off a fiber node is still a long way from fiber-to-the-home.
MCN: Today, Cox is serving 650 homes per node on average. How many homes per node will be required when bundled voice, video and data services start to achieve scale and, eventually, critical mass? What number do you foresee to handle that?
We're doing that today. Of course, you could argue that we haven't achieved mass. We've got 10 percent penetration of the homes taking a digital-video product. We've got 6 percent taking the Internet product and probably 10 percent, on average, taking the telephony product.
I think the node sizes we have today will serve us well into the future with higher penetrations. We know using 10 megahertz of bandwidth, upstream and downstream, can serve a 40 percent penetration of telephony. We've got a couple channels of data. We're at 6 percent, but we think we can probably get 15 to 20 percent for Internet services. So, I'm still looking for the application that forces me to split nodes.
Maybe one of three things would force me to do that: a heavy penetration of video-on-demand, streaming video over the Internet or video telephony. We don't have video telephony in our plant yet, and streaming video is not yet a major reality.
MCN: What are some of your thoughts about the merits involved in AT & T Broadband's testing of LightWire architecture with mini-fiber nodes that serve between 75 and 100 homes. Is Cox considering that architecture, as well?
No. From an engineering standpoint, I love fiber-to-the-home. Pushing fiber down to 100 home nodes or less? I don't think it's needed yet, but we need to be able to get there if we have to.
I'm not picking on AT & T by any stretch, but doing trials are easy. I watch trials all the time. If they launch it nationwide, I'll pay a lot more attention to it.
We're basically 70 percent done with our upgrades. We might consider 500 homes in some systems. In fact, we're going to do 500-home nodes in our recent acquisition of Northern Virginia's Fairfax County. We think the economics justify that. But 100 homes? We're not going to do that.