To Fight, Verizon Switches
Verizon’s future starts where the scuffed green floor ends on the second story of a windowless brick blockhouse in Baldwin Park, CA, a Latino community east of Los Angeles. Past that point, the building’s old asbestos-filled floor tiles were removed in March and replaced with shiny white ones, marking the place where circuit switching, the method long used to connect one phone to any other, gives way to packet switching, the technology that makes the Internet so powerful. It’s a project that promises to change Verizon’s business – and eventually, the way we all think about phone service.
Pamela and Grant Jacoby, husband-and-wife members of the Verizon project team who are showing off their new system to a reporter, say they celebrated their anniversary by going to dinner and then coming here to watch a HazMat team rip out the old flooring. Romantic? No. “But how many times do you get to see that?” says Grant, as his wife laughs.
The reason for the Jacobys’ excitement is simple: in Grant’s words, “We’re getting rid of the public switched telephone network.” But what does Verizon, the nation’s biggest phone company, build in its place? A new network that makes more efficient use of its switching stations and physical wires by working more like the Internet, and that wires up customers’ homes with high-capacity fiber-optic lines. With such an infrastructure in place, the theory holds, the longtime supplier of plain old telephone service can change into a new kind of company, one that can compete in a world where media giants like Comcast are blending services such as television, telephone, and Internet access.
Listen to people like the Jacobys, and it’s easy to imagine that in a few years Verizon customers may not even have phones, or at least not ones that only make phone calls. Instead, they’ll have devices that surf the Web, transmit video phone calls and still pictures, and deliver TV programming, TiVo style. Customers will use their cell phones to instant-message their kids’ TV screens that it’s time to stop playing video games and start doing homework. They’ll control what their phones – or rather, their personal telecommunications networks – do in a way that’s simply impossible today.
Or at least, that’s what Verizon executives hope. The company astonished Wall Street and telecom insiders in January when it announced that it would spend $2 billion over the next two years to move to Internet-type switching – a far more ambitious overhaul than those planned by its sibling phone companies, such as SBC and BellSouth. (Verizon’s changeover is distinct from the other hot trend in telephony, voice-over-Internet-Protocol services; those run primarily on the Internet, not the phone network.) But while projects like the Baldwin Park conversion are well under way, Verizon managers admit they’re spending the money without a full understanding of how the new network will be used or what services consumers will want most. According to Paul Lacouture, president of Verizon’s network services group, it’s an investment the company has to make simply to survive in a fast-changing industry. Moving to packet switching now, he says, means “we future-proof our network.”
The Paranoid Undertaker
It would be easy to mistake the Baldwin Park central office for a museum of the history of telephony: the building has hosted three generations of phone network technology since its construction early in the last century. The first and longest-lived technology was the step switch. Legend says that its inventor, a Kansas City mortician named Almon B. Strowger, was convinced that switchboard operators were sending his customers to rivals. So he built an electromechanical device that automatically routed calls according to the numbers dialed by customers and patented it in 1891. For decades, step switches occupied thousands of central offices, whirring, clicking, and banging as they opened circuits and shunted calls to their destinations.
In Baldwin Park, the mechanical step switches were finally ripped out in the 1970s and replaced with digital switches. All that’s left of the old apparatus are two walls of steel coils big enough to hold wine bottles. Now 11 rows of digital circuit-switching equipment house thousands of playing-card-sized circuit boards. Copper wires running from homes and businesses across Baldwin Park and parts of adjacent communities terminate here; to the phone network, these circuit boards are you, me, and our neighbors and employers.
In circuit-switched networks like Baldwin Park’s, a given phone call has exclusive access to a circuit, or phone line, for its entire duration. “The minute you dial your number…you’ve tied up a certain amount of bandwidth,” says Stuart Elby, Verizon’s vice president of network architecture and enterprise technologies. “So you waste a lot of space.” But the share of phone network capacity devoted to voice calls is declining fast, while data traffic is increasing, and today’s third wave of change – the transition to packet switching – is a response to that shift.
Packets are dollops of information labeled with addresses so that a computer network can forward them, node by node, to their intended recipients. As long as each packet is reunited with its neighbors at its destination, it doesn’t matter how it gets there – which means that billions of packets with different destinations can travel through the same physical wires at the same time. Though it’s the very foundation of the Internet, packet switching is a relatively old technology, first developed in the 1960s. But turning voices into packets is relatively new, and until recently the audio quality of packet-switched calls lagged far behind that of circuit-switched calls. As recently as three years ago, most engineers at traditional phone companies still sneered at packet switching. “Saying that Internet Protocol was going to be the most important protocol – well, that could be a career-limiting move,” laughs Elby.
No one’s sneering now. Verizon’s biggest fear is missing the opportunity presented by packet switching, which cable companies like Comcast are quickly seizing upon to offer voice and data services in addition to their traditional TV packages. Packet-switched networks are not only more capacious; they’re also cheaper to run. A single 100-centimeter-by-60-centimeter “softswitch” – so called because it uses software commands to handle both Internet-type packets and the older data formats used for many voice calls – accommodates as many phone lines as a 10-meter row of digital circuit switches. That saves on maintenance, and even better, it means Verizon can combine central offices: another California office, for instance, will be managed remotely by software running at Baldwin Park.
Verizon set a highly ambitious schedule for the first rollout of its new packet-switching equipment. The southern-California team assembled at the Baldwin Park office one day in early August gripes good-naturedly about the pace. Lead engineer Bill McClure says that during a conference call in February, he and fellow lead engineer Curtis E. Reese learned that they’d have until July 2 to refit five sites that handled a total of 100,000 telephone lines – a shock to engineers accustomed to testing small things for 18 months apiece. “When they said that, Curt passed out,” jokes McClure, a 32-year veteran of the company. “And then I passed out.”
“And then they told me, and I passed out,” chimes in Pamela Jacoby, the project leader. They recovered fast: this is, after all, the kind of assignment that caps a network engineer’s career. But that doesn’t make it easy. There have been plenty of double shifts and unexpected bugs. Baldwin Park is one of the first and biggest central offices to convert to packet switching, and impromptu conference calls between Verizon engineers and their counterparts at telecom equipment supplier Nortel can happen several times a day.
Getting the new equipment to work with older ISDN technology was one stubborn problem. It also showed what it will take to integrate packet-switching with the old phone network, which Verizon can’t simply turn off. Many retail and banking customers still send credit card and other transactions over ISDN connections, a 1990s predecessor to DSL. But the information traveling on one of the ISDN data channels wouldn’t transfer over to the packet switches. “That took three or four weeks to figure out,” Reese says.
The team missed the July 2 deadline, largely because of state regulations requiring phone companies to give 30 days’ notice of equipment changes. But the delay gave the team extra time to resolve technical problems like the ISDN bug. Verizon’s first fully packet-switched central office went live in mid-September; at press time, however, the changeover at Baldwin Park was temporarily on hold while Verizon attempted to resolve related legal issues.
Those first five stations handle only a small fraction of Verizon’s customers in the region, but the pace of the rollout is expected to increase, and the company’s entire network could be converted in as little as five years. It’ll be a while before customers actually see a difference in services; Verizon won’t become an amalgamated phone-cable-broadband company with gee-whiz packet-switched services until it actually has a good base of switches in place. But already, thanks to a parallel effort at Verizon to string fiber-optic cables from the curbside all the way into people’s living rooms, Verizon customers in places like Keller, TX, Tampa, FL, and Huntington Beach, CA, have spiffy new “optical network terminals” pumping up to four phone lines, television service, and 30 megabits per second of data into their homes (DSL tops out at about three megabits per second).
On the Offensive
Baldwin Park is 4,500 kilometers and three time zones away from Verizon headquarters in New York. But creating a packet-switched network that traverses that distance is the goal of Paul Lacouture (pronounced LACK-uh-chur), the executive responsible for executing Verizon’s grand plan. Lacouture inhabits a 39th-floor office that once featured a view of the World Trade Center, and his desk is the size of a conference table. But engineers who work for Lacouture say they sense a kindred soul. “It seems like he’s turned a few nuts and bolts in his time. You don’t often see that in upper management,” says Pamela Jacoby.
Indeed, Lacouture started out as a network engineer 31 years ago, at the old New England Telephone (which became part of Nynex, which became part of Bell Atlantic, which merged with GTE in 2000 to become Verizon). And the ex-engineer shares the California team’s excitement about the shift to a packet network. “It’s something that you want to do” as an engineer, he says.
But it also makes business sense. Like all traditional phone companies, Verizon faces enormous competitive pressure. Cellular phones, cable telephony, and voice-over-Internet services like Vonage and Skype are steadily chipping away at its base of residential customers (see “Skype beyond the Hype,” TR June 2004). Even more alarming, from Verizon’s perspective, is a gradual drop-off in minutes used per line. Yet at the same time, Verizon’s broadband and data businesses are booming. When, in late 2003, the cost of installing packet-switching equipment finally dropped to levels comparable to those of traditional circuit-switching equipment, Lacouture realized that it was time to embrace a new future. “The information age is really just beginning,” he says. “We’ll be in a position to package wireless, voice, data, and video. It’s a tough transition, but we think we can grow there.” After years of simply trying to preserve the company’s market share against competitive forces, Lacouture says, “this was a chance for us to be on the offensive.”
Still, it is a difficult strategy. Customers won’t be able to take advantage of packet switching for services like full-motion video unless they have true broadband connections, and Verizon believes that means running fiber-optic cable to each home or business that wants it – a costly proposition. The central-office upgrades must continue, of course, and though much of Verizon’s backbone network already uses packet switches, it’s upgrading those, as well: the company is the first buyer of Lucent Technologies’ Lambda Xtreme switches, high-end optical switches each capable of handling 1.3 to 2.6 terabits of data per second. This enhancement, while not as costly as fiber to the customer, is also not cheap.
In fact, Verizon’s upgrades will cost enough that its siblings have opted for more limited plans. BellSouth, for example, believes that technologies like DSL and Ethernet can deliver data to homes at speeds sufficient for television; it’s laying fiber-optic cables to the curbside in many neighborhoods and switching to copper wires for the last 100 meters or so. SBC is doing the same in its existing service areas, reserving “fiber to the premises” for new housing projects. Lacouture, however, remains convinced that copper is a dead end. He believes these companies will eventually wind up doing the same infrastructure work Verizon is – and by then, he says, Verizon will have a huge head start.
But bets on media “convergence” have seriously damaged other companies, including Vivendi Universal and (ominously) AT&T. In the late 1990s and 2000, under then CEO Michael Armstrong, AT&T spent more than $100 billion buying McCaw Cellular and cable providers TCI and MediaOne in order to offer local-phone and TV service and position itself for a broadband future. AT&T overspent, and both the technology and consumer demand moved more slowly than expected, which left the company vulnerable when recession hit. AT&T sold off its broadband business, has steadily lost market share, and recently abandoned its pursuit of the consumer long-distance market.
AT&T actually had the right idea, argues veteran telecommunications consultant Frank Dzubeck, but realizing it would have required technologies that didn’t yet exist. “The implosion of the industry killed [Armstrong’s] vision,” he says. Verizon is a different animal: it’s a $67 billion behemoth with an established grip on local service, the third-biggest U.S. long-distance company, and in partnership with Vodafone, the biggest U.S. cellular company. Put it all together, Dzubeck says, and Verizon has everything it needs to kill itself – and then raise itself from the dead.
Customers in Control
At least Lacouture doesn’t have to worry about whether the technology will work. Engineers at Verizon’s laboratories, especially its huge research quadrangle in Waltham, MA, have spent more than a decade looking at packet switching and know it can deliver. “We’re a reality check and a stability check,” says William C. Uliasz, director of Verizon’s optical transport network architecture group in Waltham.
The Waltham lab is where engineers developed a 12,000-page set of requirements for packet switches, to help manufacturers like Nortel develop products. It’s here that researchers attempt to bog down the fastest long-distance network on the planet – 3,000 kilometers of fiber rolled up into a couple dozen spools. And it’s here that software developers created the Universal Media Communicator, a program for desktop PCs that lets users seamlessly transfer calls from cell phones to wireless PDAs to traditional landlines to IP phones, without ever putting anyone on hold. One of the program’s myriad features represents separate phone calls as icons that users can drag and drop to initiate conference calls. The idea is to let the customers control their phone service – where calls reach them, say, or when and whether their phones ring.
It’s unclear how much of the Universal Media Communicator will show up in actual Verizon products; for now, the company is using it to inspire programmers at conventional software companies, including Microsoft, to develop applications for the Verizon network, says Michael Weintraub, director of converged services at the lab. “A lot of people think you want to blow up the public switched telephone network. But what you really want is this evolution” to Internet Protocol, Weintraub says. It isn’t that the current network is so bad – it’s just that packet-switched networks make everything faster, cheaper, and easier to use. And they turn the tables on who has control.
But will people pay for the new features enabled by packet switching? Come to think of it, what will people pay for, once voice calling is simply one entrée on a vast menu of potential services?
That’s still the big unknown. The phone companies are famously wrong about when customers will want things. AT&T first demonstrated the videophone in the 1960s. Ten years ago, phone companies thought ISDN was the path to broadband. Then DSL was supposed to be the high-speed enabler for video on demand but became mainly a way to browse the Web faster.
“We always guess wrong,” shrugs Verizon’s Elby. “So we know that we won’t come up with the killer app – but we’ll have the network to support it.”
Geoffrey Hinton tells us why he’s now scared of the tech he helped build
“I have suddenly switched my views on whether these things are going to be more intelligent than us.”
ChatGPT is going to change education, not destroy it
The narrative around cheating students doesn’t tell the whole story. Meet the teachers who think generative AI could actually make learning better.
Meet the people who use Notion to plan their whole lives
The workplace tool’s appeal extends far beyond organizing work projects. Many users find it’s just as useful for managing their free time.
Learning to code isn’t enough
Historically, learn-to-code efforts have provided opportunities for the few, but new efforts are aiming to be inclusive.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.