Thank You for Being Late

by Thomas L. Friedman

  “Once you no longer had to break up the laser light signal to amplify it, the speed that you could transmit information was no longer limited by the properties and constraints of electricity but only by the properties of light,” he explained. “Then us laser guys really got to do cool stuff.” They found all sorts of new ways to push more information using lasers and glass. These included time division multiplexing—turning the light on and off, or pulsing the lasers to create more capacity. And it included wavelength division multiplexing, using different colors of light to carry different phone conversations at once—and then combinations of the two.

  They are not done accelerating. “The history of the last twenty years is that we just keep finding faster, better ways to divide the different properties of light to pack ever more information,” said Bucksbaum. “The rate of data transfer for an undersea cable today is now trillions of bits per second.” At some point, you end up “bumping up against the laws of physics,” he added, but we are not there yet. Companies are now experimenting not just with ways to change the pulse or the color of light to create more capacity, but also with new ways of shaping that light that can deliver more than one hundred trillion bits per second down their fiber lines.

  “We are getting closer and closer to being able to transmit a nearly infinite amount of information at near zero cost—these are the kind of nonlinear accelerations you’re talking about,” said Bucksbaum. Most people right now are using this new power to stream movies, but it will infuse itself everywhere. “I ordered a book this morning at five a.m. and it’s going to be delivered by Amazon today.”

  The AT&T Gamble

  As powerful as all those fiber-optic landlines and sea cables are, they are still only one part of the connectivity story. To unleash the power of the mobile phone revolution, it was also necessary to expand the speed and reach of wireless networks.

  Many players had a hand in that, starting with AT&T and the huge bet it made that few people knew about. It happened in 2006 when the company’s COO and soon-to-be CEO, Randall Stephenson, quietly struck a deal with Steve Jobs for AT&T to be the exclusive service provider in the United States for this new thing called the iPhone. Stephenson knew that this deal would stretch the capacity of AT&T’s networks, but he didn’t know the half of it. The iPhone came on so fast, and the need for capacity exploded so massively with the apps revolution, that AT&T found itself facing a monumental challenge. It had to enlarge its capacity, practically overnight, using the same basic line and wireless infrastructure it had in place. Otherwise, everyone who bought an iPhone was going to start experiencing dropped calls. AT&T’s reputation was on the line—and Jobs would not have been a happy camper if his beautiful phone kept dropping calls. To handle the problem, Stephenson turned to his chief of strategy, John Donovan, and Donovan enlisted Krish Prabhu, now president of AT&T Labs.

  Donovan picks up the story: “It’s 2006, and Apple is negotiating the service contracts for the iPhone. No one had even seen one. We decided to bet on Steve Jobs. When the phone first came out [in 2007] it had only Apple apps, and it was on a 2G network. So it had a very small straw, but it worked because people only wanted to do a few apps that came with the phone.” But then Jobs decided to open up the iPhone, as the venture capitalist John Doerr had suggested, to app developers everywhere.

  Hello, AT&T! Can you hear me now?

  “In 2008 and 2009, as the app store came on stream, the demand for data and voice just exploded—and we had the exclusive contract” to provide the bandwidth, said Donovan, “and no one anticipated the scale. Demand exploded a hundred thousand percent [over the next several years]. Imagine the Bay Bridge getting a hundred thousand percent more traffic. So we had a problem. We had a small straw that went from feeding a mouse to feeding an elephant and from a novelty device to a necessity” for everyone on the planet. Stephenson insisted AT&T offer unlimited data, text, and voice. The Europeans went the other way with more restrictive offerings. Bad move. They were left as roadkill by the stampede for unlimited data, text, and voice. Stephenson was right, but AT&T just had one problem—how to deliver on that promise of unlimited capacity without vastly expanding its infrastructure overnight, which was physically impossible.

  “Randall’s view was ‘never get in the way of demand,’” said Donovan. Accept it, embrace it, but figure out how to satisfy it fast before the brand gets killed by dropped calls. No one in the public knew this was going on, but it was a bet-the-business moment for AT&T, and Jobs was watching every step from Apple headquarters.

  “We were expected to deal with some exponentials,” said Donovan. “And I knew that I could not get there with Moore’s law on the hardware alone. It would take too long to deploy at that scale. I had to get a faster solution—hence software. We pioneered software-enabled networking. We put everyone in the company we could muster into software development and we went to our [infrastructure] vendors and told them, ‘We are moving to software.’”

  I asked Prabhu to explain software-enabled networking, which he did with a simple example: “Think of the calculator on your phone,” he said. “It creates the virtual effect of hardware—a desk calculator—by using software. Or think of the flashlight on your iPhone. That is software using the underlying hardware to create a virtual flashlight.”

  In networking, Prabhu explained, it means minting massive amounts of new capacity for transmitting data, text, and voice by taking the same networking switches, wires, chips, and cables and getting them to work better and faster by virtualizing different operations with the magic of software. The best way to understand it is to think of telephone wires as a highway, and then imagine that the only cars on this highway were self-driving vehicles controlled by computers, so they could never crash into one another. If that were the case, you could pack so many more cars on that highway, because they could drive bumper-to-bumper at one hundred miles per hour six inches apart. When you take the electric energy passing through a copper wire or fiber cable or a cellular transmitter and you apply software to that electronic signal, you can manipulate that energy in so many more ways and create so much more capacity beyond the traditional limits and safety margins built into the original hardware.

  And just as you can set up a highway with automated cars driving at a hundred miles an hour six inches apart, said Donovan, you can “take the same copper wire designed to carry a two-ringy-dingy voice phone call and make it carry eight streams of video by maximizing how the bits perform. Software adapts and learns. Hardware can’t. So, we blew apart the hardware components, and we forced everyone to think anew. We basically turned the hardware into a commodity and then created a baseline operating system for every router, and called it ONOS, for Open Network Operating System.” Users could write programs on it to keep improving the performance.

  Software, concluded Donovan, “has power and flexibility greater than anything materials can offer. Software better captures new wisdom than materials.” Basically what we have done “is amplify Moore’s law with software. Moore’s law was viewed as the magic carpet we were riding, and then we discovered we could use software and literally accelerate Moore’s law.”

  Irwin: The Cell Phone Guy

  It was wonderful for consumers for all these networking breakthroughs to occur, but someone had to pack them into a phone you could carry in your pocket to get the full frontal revolution—and no individual was more responsible for this mobile phone revolution than Irwin Jacobs. In the pantheon of the great innovators who launched the Internet age—Bill Gates, Paul Allen, Steve Jobs, Gordon Moore, Bob Noyce, Michael Dell, Jeff Bezos, Marc Andreessen, Andy Grove, Vint Cerf, Bob Kahn, Larry Page, Sergey Brin, and Mark Zuckerberg—save a few lines for Irwin Jacobs, and add Qualcomm to the list of important companies you’ve barely heard of.

  Qualcomm is to mobile phones what Intel and Microsoft together were to desktops and laptops—the primary inventor, designer, and manufacturer of the microchips and software that run handheld smartphones a
nd tablets. And all you have to do is walk through Qualcomm’s museum at its San Diego headquarters and see its first mobile phone—basically a small suitcase with a phone on it made in 1988—to appreciate the Moore’s law journey it’s been on. Because Qualcomm today does not sell its products to consumers, only to phone manufacturers and service providers, most people don’t know about Jacobs and the role he played in launching mobile telephony. It’s worth a short reprise.

  As he explained to me in an interview in the coffee shop in the lobby of Qualcomm headquarters, Jacobs had and still has one overriding goal in life: “I want everyone on the planet to have their own phone number.”

  Now eighty-two years old, Jacobs still has that steely stubborn streak, disguised by a grandfatherly smile and warm demeanor, which is common to great innovators whom people initially dismissed as crazy: It is so great to meet you—now get out of the way while I disrupt your whole business. Oh, and have a nice day!

  We forget today that thinking you could get a phone in the palm of everyone’s hand—with their own unique phone number—back in the 1980s was not exactly an everyday dream. And it was certainly not as inevitable as it now feels. Jacobs had been an engineering professor at MIT, where he coauthored a textbook on digital communications. In 1966, he was lured west by the great weather to take up a position at the University of California, San Diego. Soon after he got there he created a telecom consulting start-up with some colleagues, called Linkabit, which opened in 1968, and which he later sold.

  In the 1980s, the mobile phone business was just emerging. The first generation, or 1G phones, were analog devices that received and transmitted over FM radio. Each country developed its own standards, and for a place like Europe—the original leader in this technology—that made it hard to roam from country to country. The next generation, 2G phones, were based on the emerging European standard for digital cellular networks, which was called GSM (Global System for Mobile) and used TDMA (Time Division Multiple Access) as its communication protocol. All European common market governments mandated the GSM standard in 1987, enabling users to roam and use their phones, and receive calls, in any western European country. The EU then tried to lobby the rest of the world to use that standard, propelled by European companies such as Ericsson and Nokia.

  Around the time all that was happening, in 1985, Jacobs and his colleagues founded a new telecom start-up called Qualcomm. One of their first customers was Hughes Aircraft. “Hughes Aircraft had approached us with a project,” recalled Jacobs. “They had submitted a proposal to the FCC for a mobile satellite communication system and they came to Qualcomm and asked if there were any technical improvements we could come up with for their proposal.”

  On the basis of his previous research, Jacobs thought that a protocol called Code Division Multiple Access, or CDMA, might be the best way to move forward, because it could vastly increase wireless capacity and therefore make mobile telephony available to many more people—and support more subscribers per satellite—than the TDMA protocol being mandated in Europe.

  At the time, though, Europe’s GSM and its TDMA-based U.S. equivalents were in their initial growth phase, and almost every investor asked Jacobs the same question: “Why do we need another wireless technology when GSM and TDMA seem good enough?”

  Both CDMA and TDMA, Jacobs explained, worked by sending multiple conversations over a single radio wave. CDMA, however, could also take advantage of natural pauses in the way people speak to allow more conversations simultaneously. This is known as “spread spectrum,” whereby each call is assigned a code that is scrambled over a wide frequency spectrum and then reconstructed at the receiving end, thus allowing multiple users to occupy the same spectrum simultaneously, using very complicated software coding and other techniques. Spread spectrum reduces the interference generated by other conversations from other cell sites. With TDMA, by contrast, each phone call took up its own slot. That limited its ability to scale, because eventually a mobile network operator would run out of slots, if too many people tried to make calls at the same time. Every network can overload, but TDMA would overload sooner with many fewer users. All in all, CDMA promised much more efficient use of the spectrum—later, it would also support the transmission of broadband data over wireless networks. In short, TDMA was the key to a finite room. CDMA was the key to an almost unlimited room. And Jacobs had an inkling that might be very important one day.

  Jacobs and his colleagues, back in the Linkabit days, had worked on one of the three networks that participated in the first demonstration of the Internet in 1977. So he could already imagine that one day cellular phones might be used to connect to the Internet. When Jacobs and his colleague Klein Gilhousen floated their alternative approach, the phone industry said it was too complicated and too expensive, and might not yield the additional capacity. And furthermore, in the early 1990s, how many people thought you would use your cell phone to access the Internet? People were just happy if their calls weren’t dropped. Meanwhile, Hughes scrapped their project with him and let Qualcomm, then an infant start-up, keep the intellectual property and patents they had developed for mobile telephony.

  Bad move by Hughes—because Jacobs would not give up.

  “So we issued the interim standard for CDMA in the summer of 1993, and we could not convince any other handset maker to make a CDMA phone,” said Jacobs. “We made the chips, software, phones, and bay station infrastructure all by ourselves—because no one else would.” In September 1995, however, Jacobs persuaded the Hong Kong phone company Hutchison Telecom to adopt Qualcomm’s CDMA protocol and phones, making it the world’s first big commercial operator of this technology.

  “Before then, everyone was very skeptical that CDMA could work in a commercial setting,” he said. “That was October 1995, and in 1996 South Korea came on using our phones made in San Diego. The voice quality was better, there were fewer dropped calls, and it could carry both voice and data at a scale that TDMA could not.”

  And that set the stage for a decisive struggle between the CDMA and TDMA protocols. While 2G phones did voice and a little text, as the popularity of the Internet grew, operators and manufacturers recognized the need for efficient wireless access to the Internet and therefore proposed a third generation, or 3G, of cellular communications that would enable you to transmit large amounts of data and voice efficiently. There was a global phone war over whose standard would prevail.

  The short story is that Jacobs won and the European GSM/TDMA-based standard lost. They lost because their technology had a finite amount of spectrum and CDMA enabled you to do so much more with the same amount of spectrum—and there was soon so much more to carry, thanks to the Internet. We don’t remember these wars today, but they were bloody. The U.S.-invented standard prevailed not only because it was better but also because, unlike in Europe, where the governments mandated a standard, in the United States the government allowed the market to choose, and many chose the Jacobs CDMA pathway. Again, you probably missed most of this, but it had huge implications. The vast majority of the world’s population when they access the Internet today do it through a phone and not a laptop or desktop. And the reason that happened at the speed and price that it did—making smartphones the fastest-growing technology platform in history—was Jacobs’s early recognition that CDMA would efficiently support Internet access as well as voice.

  Sure, you could say that in the end everything gets invented and someone would have found their way to CDMA as the foundation of mobile Internet. Perhaps. But it was due to Jacobs’s titanic stubbornness in pushing the CDMA standard, when no one else thought it necessary and Europe was pushing the other way, that it happened faster and farther and cheaper. And one by-product was that American phone companies took the lead on 3G and 4G. Meanwhile, once its protocol and software were taken up for mass adoption, Qualcomm got out of the business of making phones and transmission platforms and just focused on the chips and software.

  Today, said Jacobs, “people ev
erywhere in the world have both voice and efficient access to the Internet and that supports education, economic growth, health, and good governance.” One key “reason we won,” he added, “was that even though CDMA was more complicated to implement, people were just thinking about the capacity of chips at that moment in time. They were not taking into account Moore’s law that would allow the technology to improve every two years and enable the greater efficiency that could be achieved with CDMA.” People say that in hockey you don’t go where the puck is, you go where the puck is going, and Qualcomm went where the puck was going: to Moore’s law, which was on a hockey stick–like curve upward. “Somewhere in the early 2000s when we were trying to expand to India and China,” said Jacobs, “I made the outlandish prediction that one day we would see hundred-dollar phones. Now they’re below thirty dollars in India.”

  But the Jacobs family inventions did not stop there. In late 1997, Paul Jacobs, who later succeeded his father as CEO, had a brainstorm. One day he came into a staff meeting in San Diego, took a Qualcomm cell phone and taped it together with a Palm Pilot, and told his team: “This is what we’re going to do.” The idea was to try to create a device that combined the Palm Pilot—at that time basically a combination calendar, Filofax, address book, and day planner, with note-taking capabilities and a wireless Web-based text browser—with a 3G cell phone. That way when you called up a phone number in the Palm Pilot address book, you could just click on it and the cell phone would dial it. And with the same device you could surf the Internet. Jacobs approached Apple to see if they were interested in partnering with Qualcomm on this, using the Apple Newton, their Palm competitor.

  But Apple—this was just before Steve Jobs came back—turned them down and eventually killed the Newton. So Jacobs went to Palm and together they ended up making the first “smartphone”—the Qualcomm pdQ 1900—in 1998. It was the first phone designed not just to relay text messages, but to combine digital wireless mobile broadband connectivity to the Internet with a touchscreen and an open operating system that eventually ran downloadable apps. Qualcomm later created the first mobile telephone–based app store, called Brew, which was marketed by Verizon in 2001.