Finally, AT&T wanted to make money at this game—this is the telephone company, after all, not Mother Teresa—so it needed a way to bill customers. In the old days, when operators were manually switching calls, this was easy: long-distance operators wrote up a paper toll ticket for each call. These tickets were collated and processed by hand. But if machines are doing the switching and routing, machines need to be able to do the billing too. It wouldn’t do to have an automated network that could handle millions of calls a day only to have the entire operation bog down because humans had to tally up the bills by hand.
All of this was an incredibly tall order in the 1930s. Yet the crazy thing is Bell Labs got right to work. It would take tens of years, thousands of engineers, millions of dollars, and buildings full of equipment to make it happen. In the end the telephone network would be transformed into something previously undreamed of: it would become the largest machine in the world, one that would eventually extend over the entire surface of the earth.
Perhaps the best way to follow this transformation is to start by putting your finger into the hole marked 7 on an old-school rotary telephone, maybe back around 1950 or so. Crank the dial —the actual metal dial—all the way over to the right until your finger is up against the dial stop. Remove your finger. A spring unwinds, spinning the dial back to the left. As it spins, over the course of about three-quarters of a second, your telephone sends seven electrical pulses down your telephone line, over the wires and cables in your neighborhood, and into one of several hundred Strowger switches in your local central office.
In movie terms, that Strowger switch was Frankenstein’s monster writ small: able to follow simple commands—the simpler the better —but a little short on brains. Each Strowger “can” was a cylinder about sixteen inches high and about six inches in diameter, jam-packed with wipers and ratchets and pawls and blades and other mechanical clockwork. Your telephone dial directly controlled its musculo-skeletal system. Every one of those electrical pulses your phone sent down the wire made something twitch inside the Strowger switch it was connected to. “Twitch,” by the way, is not figurative; it is an accurate description of what physically took place in the switch. The digit 7 that you dialed caused a pair of metal contacts to twitch upward seven times, so fast that it seemed to make a brrrp noise as it went. While waiting for your next digit, it rotated to the right, connecting your telephone line to the next idle Strowger switch it could find. That next switch would then accept whatever digit you dialed next, again twitching a mechanism inside up and to the right. This mechanized ballet continued until you had dialed all the digits of your number; your last digit connected you from the last Strowger switch in the switching train to the actual pair of wires running to the telephone you wanted to call.
The key thing about the Strowger system was that every pulse your telephone sent down the line caused something to happen in the switch—physically, immediately, and directly. This direct control system was innovative when it was invented in the 1890s. But in addition to having lots of noisy, moving parts that needed service and eventually wore out, it suffered from two fundamental problems. First, every digit you dialed tied up one Strowger switch for the duration of your telephone call. If telephone numbers in your local exchange were four digits long, then when you called your friend down the street and talked for an hour, you tied up four Strowger cans for the entire call. This meant the telephone company needed to cram a lot of these Strowger switches into a central office, and that was expensive.
The other problem was that, like Frankenstein’s monster, Strowger switches were not the sharpest knives in the switching drawer. Because calls proceeded through a step-by-step switching system one digit—and one switch—at a time, no individual Strowger switch ever saw more than a single digit of the telephone number you were dialing. Nothing in a Strowger system had the big picture, and that limited what the telephone system could do.
The wizards of Bell Laboratories gave telephone switches a brain of sorts when they developed the successors to the Strowger switch. Both the panel and crossbar switching systems used a technology that the telephone company called common control. Instead of the telephone directly controlling a switching machine itself, your telephone would tell the switching system’s brain what you wanted done and the brain would figure out how to do it. So, for example, in a crossbar central office, the digits you dialed on your telephone no longer caused the central office’s switching system to twitch directly with every pulse your phone sent out. Instead, your digits were stored in a relay-based memory called a sender. Once you had dialed the full number, the switch’s brain, the marker, could look at the digits and figure out what it needed to do to connect your call. Once it did this, it could forget about your call and move on to the next one, freeing up resources. And because the brain had the entire telephone number you wanted to dial in one convenient place, it could do clever tricks that a Strowger switch could only dream of. As Bell Labs’ head of switching later wrote, “In a word, the [switching] systems were acquiring a form of machine intelligence.”
The pinnacle of that era’s telephonic mechanized brain was something called the #4A crossbar switch. It was another in a long line of creatively named products from AT&T, joining the ranks of the #1 manual switchboard, the #5 manual switchboard, the 500-series desk telephone, and the #1 crossbar switch. Who needs fancy product names when you’re a government-sanctioned monopoly?
For what it was, the 4A deserved a grander name. Deployed in 1950, it was a triumph of common control switching, the most advanced switching machine created to that point in history. Even the word machine doesn’t do it justice: it conjures up images of a mechanical contrivance, something bigger than a breadbox but smaller than a car; a lawn mower, maybe. In contrast, the 4A took up a good chunk of a city block. Built up of rack after rack of gray metal cabinets filled with crossbar switches, wiring frames, markers, senders, and relays, if the Strowger switch was Frankenstein’s monster, the 4A was Godzilla. By 1960 there were fifty-nine of them throughout the United States; almost two hundred of these behemoths would eventually be installed, the last in 1976.
The 4A was to be the brains of the long-distance network, the magic switching machine that could automatically figure out how to route a long-distance call from one place to another. Its routing intelligence did not come from a computer but rather from a device called a card translator. Hundreds of thin steel cards, each about five inches wide and ten inches long, had patterns of 181 holes punched in them to indicate how a call should be routed. Based on the first six digits of a telephone number—the area code and the exchange number—electromagnets selected and dropped cards. Light shined through the holes. By seeing where light passed through and where it was blocked, the 4A could decide how and where to send the call as well as figure an alternate route if something went wrong with the first one. As the telephone network grew and changed, the 4A could be reprogrammed simply by changing out cards. Even if the 4A fell short of human intelligence, the telephone company knew that its common control systems were nothing to sniff at. “At the end of this era,” wrote a former Bell Labs executive, “Bell engineers were able to look back on the automated network of switching systems as the largest distributed computer in world.”
Just like human operators, the brainy 4A switches passed calls among themselves and their less intelligent brethren by talking to each other. And like human operators, there were only a handful of things the switching machines needed to tell each other: what number to dial, whether the called party answered, and whether either party hung up. The telephone company called these latter two items supervisory information, since they had to do with how an operator would supervise a call. They were critically important: you can’t charge a customer for a call if you don’t know that the call was answered or when the parties hung up.
AT&T enabled its long-distance telephone switching machines to talk to each other by teaching them two
different signaling languages: single frequency and multifrequency. Both were based on the switching machines sending tones—musical notes, basically—down the telephone trunk lines to each other. The multifrequency language, or MF for short, used pairs of tones to communicate what digits to dial, much the same way that today you use touch tones to communicate to the telephone system what digits you want to dial when you make a call from a landline telephone. The other language, single frequency or SF, was simpler than MF, and although it was slower it could be used with less intelligent switching machines, such as the old step-by-step switches. SF used pulses of a single tone—2,600 Hz, or seventh octave E for the musically inclined—to communicate dialing information: one beep to dial a 1, two beeps to dial a 2, etc. In a sense, it was just like a rotary phone sending electrical pulses down a phone line, except that it sent beeps instead. Both SF and MF also used this 2,600 Hz tone for supervisory information, that is, to communicate when one machine wanted to make a call and when the person you were calling answered the phone so that billing should start.
Introduced in the 1940s, MF and SF were high tech for their time. The multifrequency system was speedy, taking only a second or so to transmit a ten-digit telephone number from one switch to another. The tones sounded like fleeting musical notes and customers could sometimes hear the quick little blips of MF digits as they waited for their calls to go through. AT&T began acting like a proud parent of a musically gifted child. Magazine ads in 1950 showed a musical scale with the pairs of notes that made up each MF digit and described the system as “playing a tune for a telephone number.” Telephone bill inserts bragged about MF and the tones were featured in an educational AT&T movie as well. In a flight of fancy, one telephone company manager told the press that new AT&T switching machines “sing” to each other.
The cleverest thing about SF and MF signaling was this: they allowed the switching machines to communicate by using the exact same wires that humans used to talk to each other. AT&T had spent millions of dollars running long-distance cables all across the United States. These cables were designed to carry voice, since that’s what AT&T’s human customers and operators used speak to one another. Instead of building a separate computer network for its switching machines, AT&T realized it could reuse its existing long-distance telephone circuits to carry both human voice and signaling information for each call. This would cost less than building a separate network and would be faster to deploy. This approach, called in-band signaling, meant that signaling information was sent in the same frequency band and over the same wires that were used for voice. It was an elegant and economical solution to the problem.
With the crossbar switch and the multifrequency signaling system, AT&T could embark on the next phase of automating long-distance switching, something called operator distance dialing. The idea here was to allow operators to directly dial long-distance calls, even if customers couldn’t. If you wanted to call coast to coast, you’d still call the long-distance operator. But instead of the operator having to plug cords into jacks and talk to other operators and build up a lengthy chain of connections, circuit by circuit, she would just key in the area code and telephone number on a keypad on the console in front of her. The switching machines would do the rest, routing the call and talking to each other with MF or SF to set up the intermediate links. If the place she was calling couldn’t be reached by just keying a number into her console, she would use the machines to get her call as far across the country as she could and then enlist the help of a plug-and-jack manual inward operator who was closer to the final destination.
Operator distance dialing simplified and sped the dialing of long-distance calls—good for customers since their calls went through more quickly and good for the phone company because it needed fewer operators to handle more calls. But it also provided AT&T with an opportunity to work the kinks out of automated long-distance switching without having to directly involve its customers. The long-distance operators became the first users of the new automated long-distance network—beta testers, we’d call them today. Or, as they were called by an AT&T spokesman at the time, “guinea pigs.”
The guinea pigs survived and AT&T decided the kinks had worked out enough to let the customers try it themselves. On November 10, 1951, the small town of Englewood, New Jersey, became the first place in the country where customers could dial their own long-distance calls. Instead of dialing 211 and telling the long-distance operator they wanted “Garfield 2-2134 in San Francisco,” lucky Englewood residents instead picked up the telephone and dialed ten digits themselves: 318-GA2-2134. Their local telephone switch would take this number, find a trunk to the remote city, and then send the musical MF notes down the line to get the call across the country. In essence, the local telephone switch acted as a sort of translator, taking the digits you dialed with your rotary phone and converting them to the telephone network’s internal language of MF tones. (It worked the same way when touch-tone dialing was introduced years later: the local switch translated the touch-tone digits you dialed on your phone into MF digits that it sent into the long-distance network; this was necessary because touch-tones weren’t the same as the MF tones.) Best of all, all this happened in seconds, not the minutes that used to be required when operators were involved.
To start with, Englewoodians were able to directly dial some 11 million people in Philadelphia, Pittsburgh, Cleveland, Detroit, Chicago, Milwaukee, Oakland, San Francisco, and Sacramento. Over the next twenty years customer long-distance dialing—later known as direct distance dialing, or DDD—spread across the country, with more and more customers able to dial their own long-distance calls. The largest machine in the world was growing, and the engineers at Bell Laboratories were finally getting their wish: a fully automated long-distance network, one where calls could be dialed coast to coast without operator intervention.
It would turn out to be a classic case of that old expression “Be careful what you wish for.”
Five
Blue Box
RALPH BARCLAY WAS walking through the engineering library at Washington State College, just minding his own business, when it called out to him. He couldn’t say why, it just did.
It was a booklet, about seven by nine inches and maybe half an inch thick, on display in the library’s new periodicals section. Its pale blue cover proclaimed it to be the November 1960 issue of something called the Bell System Technical Journal. It had been out for less than a week.
Barclay looked at the table of contents printed on its cover. Most of the articles could put even the hardest of hard-core geeks to sleep at twenty paces: “Magnetic Latching Relays Using Glass Sealed Contacts,” “Molecular Structure in Crystal Aggregates of Linear Polyethylene,” and the ever popular “‘Ionic Radii,’ Spin-Orbit Coupling and the Geometrical Stability of Inorganic Complexes.”
Yet one title caught his eye: “Signaling Systems for Control of Telephone Switching.” He flipped to the article and started skimming. Minutes passed. His original purpose for coming to the library shelved for the moment, he sat down and began to read in earnest.
Barclay was just eighteen. Athletic and of medium build, with brown hair and blue eyes, Barclay had started his first year at Washington State’s Pullman campus, about fifty miles south of Spokane, just a couple of months earlier. “I was living in the dorm,” he remembers, “and a lot of people in the dorm are looking for ways to make cheap phone calls home to their girlfriends and parents and suchlike.” One of the guys in the dorm had—“somehow,” he says—acquired his own personal pay telephone. And although students weren’t allowed to have telephones installed in their rooms, for some reason the dorm rooms had telephone lines in them.
Barclay’s dorm had quite a few engineers in it, and engineers, Barclay allows, are a problem. The engineers soon determined that somebody had left the door unlocked to the building’s telephone closet, the little room where all the telephone wires come from
. In the dark of night an operation was mounted. Certain wires were cross-connected. Et voilà: a pay telephone line from somewhere on campus ended up connected to the personal pay phone in Barclay’s dorm. Barclay and the other kids in the dorm could now make telephone calls by depositing money in the pay phone, as usual, but the difference was that the owner of the pay phone—apparently not a business major—was a nice guy and returned the caller’s money after each call.
Maybe it was this pay phone hack that caused bells to ring in Barclay’s brain when he spotted the article in the Bell System Technical Journal. It laid bare the technical inner workings of AT&T’s long-distance telephone network with clarity, completeness, and detail: how the long-distance switching machines sang to each other with single-frequency (SF) and multifrequency (MF) tones, how 2,600 Hz was used to indicate whether a telephone had answered, what the frequencies were of the tones that made up the MF digits, how overseas calls were made, and it even included simplified schematic diagrams for the electrical circuits necessary to generate the tones used to control the network. It was all there. Nothing was hidden.
Exploding the Phone : The Untold Story of the Teenagers and Outlaws Who Hacked Ma Bell (9780802193759) Page 6