Steven Cherry: Hi, this is Steven Cherry for IEEE Spectrum’s “Techwise Conversations.”
The 134-year-old U.S. public phone network is dying—a third of all U.S. households are already cellphone-only. And it’s not just cellular that’s killing it. More and more businesses and households are trading their traditional switched telephone service for voice over Internet Protocol, or VoIP, services. That’s led to a paradoxical situation, where a huge number of phone calls start out as Internet packets and end as Internet packets, but have to be switched to, and then from, a voice circuit in between.
What remains is to put the Internet protocol in the middle of the network as well. And it’s happening. In a 2009 filing with the U.S. Federal Communications Commission, [PDF] AT&T, the largest phone provider in the world, indicated its support for “the transition from the circuit-switched legacy network to broadband and IP-based communications.” It went on to say, “That transition is under way already: With each passing day, more and more...services [are] leaving the public switched telephone network and plain-old telephone service as relics of a bygone era.” And the company called on the government to give a date certain for the last plain old telephone call.
In June, a Washington, D.C., advisory group, the Voice Communication Exchange Committee, formed and committed itself to a complete transition to the Internet protocol by a date of its own choosing: June 15, 2018. The changeover will provide some enormous benefits to all of us, not least of which is high-definition voice service, similar to the transition from cathode-ray television to HDTV. And it will provide some enormous benefits to the phone companies, including the surprising one of windfall profits from their legacy real estate holdings.
My guest today is the founder of the Voice Communication Exchange Committee, Daniel Berninger. He’s an old Bell Labs hand from the 1990s, where he help the old AT&T first come to grips with voice over IP. He’s also had a hand in just about all the early startups and deployments of VoIP, including Free World Dialup, VocalTec Communications, the VON Coalition, and Vonage. He joins us by, of course, telephone.
Daniel, welcome to the podcast.
Daniel Berninger: Thank you, Steven. Thank you for having me, and thank you for covering this topic.
Steven Cherry: Daniel, maybe start by describing a modern telco, whether it’s AT&T or China Telecom for that matter, and tell us where the Internet protocol has and hasn’t invaded and what still needs to be replaced.
Daniel Berninger: Sure. It’s interesting, actually, you should say “modern telco.” It’s almost an oxymoron. Telcos—the service they deliver hasn’t changed in about 80 years. So you and I are communicating via a telephone here, and we could have done that in the ’50s, and—in the most part—in the ’40s, and the only difference would be maybe a little bit different noise levels. A modern telco, however, you know, over the years, the phone companies certainly change what’s under the hood, and so they do go through multiple technology evolutions. The most recent one was to bring IP protocol into their networks. So most telephone companies these days move voice traffic and other types of traffic around their networks using Internet protocol. And so they sort of mix, in some cases, networks and parts of their network that would support data versus voice.
Steven Cherry: So how does the wireless world fit into all of this?
Daniel Berninger: You can think of wireless as just a mile or a half a mile worth of air interface, and everything else about it ends up on wires and fiber and operates in a similar way that the wireline world works. The two networks, the wireline and wireless networks, grew up separately, so they have, to some extent, separate switches and separate infrastructures. But the vast majority of the wireless network is also wires and uses very similar technology, and one of the wins that we get in the transition to all IP networks is everything will be a single network—so whether a fully unified network, whether they are moving video, or a wireless call, or a wireline call.
Steven Cherry: I guess earlier wireless protocols would have had trouble with that, but the 4G networks we’re all moving to, the LTE, is basically IP to begin with—is that right?
Daniel Berninger: Right. So LTE, which should come in the next two to three years and is out there to some extent for data already, is a native IP all the way. It’s built as a data platform, and then they overlay voice onto it. And previous to that, in the ’90s, of course, we had a voice infrastructure, which we overlaid data. But, of course, there’s much more data these days, whether voice or not. So the networks are optimized for data these days.
Steven Cherry: Some phone companies, and not just in terms of wireless, but some wired phone companies, have sort of tried this before, right? British Telecom had an ambitious program, for example, to phase out traditional telephony in favor of the Internet protocol. How did that work out?
Daniel Berninger: A number of carriers have moved along in terms of implementing Internet protocol in the middle of their networks. British Telecom did and was able to sell a lot of their real estate as the bulk of their equipment shrunk. But what happens is they still end up exchanging traffic within each other in the same old circuit-switch way. So just about any carrier—if you think about the world’s largest carrier by revenue, AT&T, they are only about 2 or 3 percent of global telecom traffic. Ultimately, just about any carrier has to send a call—any call that it gets—off its own network. So the infrastructure between networks is the key bottleneck at this point.
Steven Cherry: How does that work? In the U.K., for example, phone calls are all IP, unless they are international calls, and then basically they all go to some central point, the phone calls, and interface with the traditional telecom world?
Daniel Berninger: Right, because the regime for moving from one telephone company to another is a regulated, government-mandated regime ever since about 1913, after the Kingsbury Commitment. Governments govern how one telephone network connects to another. In all those rules are sort of built in a lot of assumptions about old infrastructure, traditional circuit switching, TDM networks, and so it’s very complicated and awkward to move off of that.
Steven Cherry: You mention that the real estate savings—I think you were the first to notice the enormous gains that the big telcos stand to make. In January, you wrote an article in which you used the word “windfall.” I guess the key to it is IP routers take up a lot less space?
Daniel Berninger: Well, it’s a couple of things: When the regime for connecting networks was assembled in the last century, it was put together in a way that would make it difficult for new telephone companies to operate, what we call “anti-competitive.” If you’re an architect and you want to design something that would make it really hard for someone to get into the telephone business, what they did was, connecting the networks was like connecting branches of a tree, so you have thousands of branches.
AT&T, for example, has something like 5000 central offices, and in order to interconnect, you have to show up at each of those offices and connect your trunks. That process of having one tree talk to another through the limbs is extremely expensive, and in the places where the networks meet, they use equipment that is 30 to 40 years old, extremely cumbersome. Some of the stuff still has discrete components in it, and essentially, when you go to Internet protocol, you can collapse by 98 percent the number of places you meet networks. A company like Comcast, which built its network in the 2000s, has five central offices, versus AT&T’s 5000. They obtain a whole lot of savings relative to their network infrastructure relative to AT&T. Eventually AT&T can obtain those same benefits. As we speak, AT&T is one of the largest commercial real estate operators in the country.
Steven Cherry: The other big telco in the U.S. is Verizon. They had an ambitious capital expansion project to replace copper lines with optical fiber, and suddenly a couple of years ago they backed off, because they were apparently investing too much money in their network and not enough money in their shareholders. Even with the real estate “windfall,” that you call it, are the big telephone companies really going to make the big investment it’s going to take to phase out the old network and make everything IP?
Daniel Berninger: I don’t want to speak for the telcos. I can give you an analyst’s perspective. Essentially the transition we’re going through as we transition to all IP networks reduces the operational expenses of what it takes to move voice or any other type of media from point A to point B. The good news is it gets a lot less expensive to do things. The second thing that happens is you’ll be able to offer new services—HD voice being a first example—but you’ll also be able to integrate applications and voice a lot more completely and seamlessly, and so the carriers will be able to increase their revenue per user. If you have a situation where you have the possibility of increasing your revenue and decreasing your cost dramatically, it’s pretty appealing for anyone to pursue, and on top of that there’s the whole dynamic about the regulatory treatment of these networks.
Steven Cherry: I wanted to get to a high definition of voice. I remember back in 2000 or so, going to Avaya Labs—this was one of the Bell Labs spinoffs—and seeing and hearing HD voice, in stereo, in fact, and yet so far, like our office phones are voice based, and I’ve used Vonage and Skype, and so far, IP has been a giant step backward in audio quality instead of a step forward. What happened, and is it going to get better?
Daniel Berninger: The reason we don’t see HD out there today is because it’s a really difficult transition to make, in that essentially, when all issues have been considered, there’s no way to move to HD unless the entire industry moves to HD. Once you have a challenge that requires an entire industry to do something, then of course you’ve got a very large challenge. The reason the whole industry has to move is because when you pick up a phone and you have 6 billion endpoints in theory you can dial, 6 billion numbers you could dial, if a particular carrier implements HD, all of a sudden, great, I have 3 or 4 percent of the potential numbers I could dial could be HD, and it’s a very difficult educational process. And for the carrier to talk to his consumers and say, “Hey, every one in 33 calls you make could be an HD call,” the value proposition is very weak unless the whole industry moves.
Steven Cherry: As I understand it, an ordinary landline phone call would use the equivalent—that’s an analog call—would use the equivalent of 60 kilobits per second, and HD would require about four times that: 256 kb.
Daniel Berninger: There’s two different things: One is frequency response the user hears. Your ear—you and I are getting older, so we can’t hear quite as well as our children—but, generally, humans can hear from about 20 cycles up to about 20 000. Today, and actually since about the mid-’30s, telephone calls provide about 300 hertz and up to 3300 Hz. That’s obviously not even half that range—the 20 to 20 000—and it’s not even a third. It’s an important aspect of the range, but there’s a whole lot of communication content that gets dropped on the floor. HD doubles that. So in HD you go from about 80 to 7000. It captures a lot more on the low end, and the presence and the feeling of who’s talking, and on the high end it captures a lot more of the high-frequency sounds that make it easier to tell the differences between the separation in words.
There’s another bandwidth issue, and that’s what it takes to move that HD call from point A to point B. There the news is good in the sense the codex does not require additional capacity. A wireline call would consume 64K and move from A to B. An HD call would require the same. In the case of wireless, where they are even more stringent on use of bandwidth, a traditional wireless voice call might use something like 12.5 kilohertz, and HD would be, like, 13.5. So it’s very little difference. It’s not a problem of network capacity.
Steven Cherry: One last question: How did you pick your date of June 15, 2018?
Daniel Berninger: We knew that we needed to pick a date certain to finish the project. If you hire an architect to build you a house, you don’t sit there and say, “You know, just let me know when it’s done.” You talk to the architect. You say, “I have a move-in date, and we’ll plan ourselves around that.” So we knew we needed a date certain. It’s the future, so people say, well, is it three years or is it 10 years? We picked six years as the duration of the project, because historically six years is enough to get quite a bit done. Both World Wars took less than six years. The Panama Canal was completed in six years. Six years is enough time to do pretty much anything humans can think of doing. So we picked six years.
As far as June 15, 2018, it just so happened we got started and got all of our companies talking together on June 15, 2012, so we just called that our start of the clock, and we were off to the races.
Steven Cherry: Well, I hope you do pull it off by your end date. It’s kind of an amazing development as far as I can see, much bigger than the transition to digital television, maybe by orders of magnitude.
Daniel Berninger: The reason it’s bigger is because television is not baked into the economy the way communication is. Obviously they sold a lot of flatscreens, and the consumer electronics industry did very well. The industry itself, the revenue doubled in the years leading up to the transition date. But in the case of communications, telephone calls, and just communications in general, they are baked into the economy, so we’ve seen great economic gains every time we get a big communication improvement, whether we’re talking about the Pony Express or telegrams, telephones, or radios, etc. This is a big paradigm shift. I would say it’s a once-per-century development.
Steven Cherry: Good luck with it. And thanks for joining us today.
Daniel Berninger: No problem. Thanks for having me.
Steven Cherry: We’ve been speaking with Daniel Berninger of the Voice Communication Exchange Committee about shutting down the last phone switch by June 2018.
For IEEE Spectrum’s “Techwise Conversations,” I’m Steven Cherry.
Announcer: Techwise Conversations is sponsored by National Instruments.
NOTE: Transcripts are created for the convenience of our readers and listeners and may not perfectly match their associated interviews and narratives. The authoritative record of IEEE Spectrum’s audio programming is the audio version.