In the early 1960s, fears of nuclear attack exposed how fragile military communications were—so a computer simulation was built to test a resilient way to route messages. The work, led by programmer Sharla Perrine Boehm while at RAND, later fed into packet swi
When the U.S.. ballistic missile early warning systems went dark overnight on November 24, 1961, the stakes were immediate and terrifying.. Officers at a base in Omaha tried to reach communications headquarters in Colorado Springs. only to find the phone lines dead.. General Thomas Power. commander in chief of the Strategic Air Command. was pulled awake—and nuclear forces were put on full alert as bomber crews guided planes onto runways.
Minutes later, communications resumed, and headquarters reported there was no attack. The outage, it turned out, traced back to a single motor at a relay station in Colorado that overheated, cutting off the one relay station that all communications passed through.
A near-accidental nuclear war—triggered by what was effectively a fragile link in the chain—left the U.S.. military with a clear demand: invent communications that could survive an attack that knocked out conventional systems.. The solution was not just theoretical.. It depended on a computer simulation created in the early 1960s on 1960s-era computers. designed to show that messages could still get through when parts of a network were damaged.
That simulation would later be recognized as foundational to packet switching and the ARPANET—an early network built by the DoD’s Advanced Research Projects Agency. eventually evolving into the modern Internet.. And yet. in the story of internet origins. the person who helped make the idea demonstrable. Sharla Perrine Boehm. often sits just outside the frame.
In the oral history recorded by the Charles Babbage Institute in 1990. computer scientist Paul Baran described the danger plainly: “there’s no communications that can survive an attack.” At RAND Corporation in Santa Monica. California. Baran and others focused on national security problems tied to a communications system vulnerable to nuclear strike.. Military communication in the early 1960s largely ran over phone lines and shortwave radio, routed through centralized circuits.. Those messages typically moved through at most five nodes. or connecting points—so if those nodes were hit. the network went down.
Baran’s conviction was to change the underlying approach. Instead of phones or radio, he envisioned messages sent through computers using a network without any central node. Each node would connect to its neighbors, and the network would remain functional even if some nodes were knocked out.
Selling that idea proved difficult.. Baran’s colleagues. as he recalled. were steeped in analog communications rather than digital processing. and they sometimes dismissed his claims as sounding like nonsense.. The resistance extended beyond RAND.. Baran also had to persuade the managers of AT&T. which controlled long-distance communication and held a monopoly in the context of what could be done with the Air Force.
When Baran pitched the concept to AT&T, the reaction came with skepticism and condescension.. He recounted engineers responding with: “Wait a minute, son.. You mean you open the switch here before the traffic has left the other end of the country?” He said he answered yes. and described the room as shaking its heads as if. “this is how a telephone works.” The tension was not only technical; it was about whether a digital network could move a message from A to B reliably without depending on central nodes.
Baran needed proof. He found it, in large part, in the work of a young math teacher across the street at Santa Monica High School: Sharla Perrine.
Even the details of her path reflect how out of place she could feel in that era.. Doug Rosenberg. a systems engineer who was a close friend of Sharla’s husband. Barry Boehm. said there weren’t women in engineering jobs in the early 1960s—“a bunch of guys in crew cuts.” But Sharla’s household. as Rosenberg described it. was shaped by necessity rather than convention: her mother immigrated to the U.S.. from Sweden and raised Sharla by
herself after her marriage ended around the time Sharla’s older sister died in 1932; Sharla was 2 at the time.. In that environment, there were no gender roles in the practical sense.. Tenley Burke. Sharla’s daughter. later described her grandmother as a carpenter who taught Sharla’s mother to fix things and create things “and not have to buy them.” Burke added that her own mother carried that same “no-nonsense. let’s do it ourselves” mindset.
Boehm’s technical story began with mathematics.. After earning a degree in teaching from UCLA. she taught math—first at a junior high and then at a high school—while also gravitating toward programming.. While waiting for her security clearance in 1959. she met Barry Boehm. and Burke later described how Sharla was drawn to RAND because it was local and filled with people thinking big thoughts.. She also liked to talk with men at parties because they were discussing “interesting things.”
In the early 1960s, her big assignment arrived.. Baran needed someone to prove his concept could work: a network resilient to catastrophic failure.. Rosenberg described the simulation approach as a method to test whether the idea could function: “The way you figure out if it’s gonna work or not is you simulate it.”
In the simulation. Boehm broke a message into small packets—an analogy laid out as cutting up a letter and sending each piece on a different route from Los Angeles to New York.. Rather than relying on a hard-wired path like a phone line, the packets would move through different sequences of nodes.. If enough neighbors remained connected, the network could still deliver the message.
Her simulation also incorporated a “damage” subroutine.. With it, she could knock out nodes—“knocking out these five nodes”—and watch how the system responded.. Rosenberg explained what mattered: packet loss was expected. but if copies of each packet traveled along different paths. at least one copy could reach the destination.. Serial numbers in packet headers then allowed the pieces to be reassembled in order.
Baran later connected this approach to what he called “hot potato routing. ” because at each node the packet had to be kicked onward quickly.. The crucial part, as the account describes it, was that there was no fixed path for each packet.. The network determined the route in real time. and when a node went down. the network shifted the packet’s route instead.. In effect, the network could “heal itself.”
Rosenberg went further on what Boehm’s simulation demonstrated: if the simulation was running and suddenly some of the five nodes were removed. the remaining nodes reorganized and communications restarted.. In Boehm’s simulation. even if half the network was destroyed instantly. the remaining nodes reorganized and got communications going again in under a second.
That speed and adaptability were central to what Baran needed. As the program recounts, Boehm’s work gave him proof the idea was possible and showable as a nationwide system with intelligent resilience.
Still, it was not the technical side alone that determined what happened next.. Baran’s frustration with AT&T never fully faded.. He said AT&T’s objections were not only technical.. He described the company saying first it wouldn’t work. and second. if it did work. they weren’t going to put the project “in competition with ourselves.” In the end. AT&T never seriously backed the effort.
But both RAND and the Air Force were willing to move forward with implementing hot potato routing in practice, even without AT&T’s dollars. The plan required funding—through the Department of Defense. Here, a different kind of uncertainty entered.
Baran did not have confidence in the Defense Communications Agency, the group that would receive funding from the DoD.. In the oral history, Baran said, “We had zero technical competence.. We had some discussions and mutually concluded that all they would do is screw it up.” He said he ended up recommending the agency not proceed. and the simulation’s work was effectively shelved.
The DoD did not give the agency the funds to build the distributed network, and Boehm’s “brilliant work got tucked away” until it reappeared years later in a different form.
That reemergence came with ARPANET.. In October 1969. a message appeared on a computer screen at Stanford University: the letters “l-o. ” sent from a computer hundreds of miles away at UCLA.. The rest of the message, meant to spell “login,” never arrived.. Even so, that incomplete attempt was described as the first communication sent over ARPANET.
Within a few years. dozens of universities and government-run research organizations connected to the network. and the messages across it relied on packet switching—tiny chunks of data sent over changing routes.. The technology was described as making it possible for chopped-up packets to travel across the network. from UCLA to Stanford or MIT or NASA. and then reassemble at the destination.
At the heart of the story was the relationship between Boehm’s simulation and what ARPANET later used. Packet switching—defined as messages being chopped into packets and sent with routes that change in real time—was also the technique “hot potato routing” ultimately became known as.
Tony Rutkowski described the internet as layered.. He said it’s “another layer on top of the digital packet network. ” an “intelligent adaptive network on top of another one. ” with the ability to “concatenate these things in different layers.” In the account. Sharla’s work is described as underpinned in that foundation.
In the 1990s, debates returned with renewed energy over who deserved the label of “father” of the Internet.. Names circulated: Paul Baran was included for his work on packet switching.. Another was the British computer scientist Donald Davies, working on similar ideas around the same time.. And Leonard Kleinrock argued that packet switching’s invention belonged to him. saying he was the “Inventor of the Internet Technology.” Boehm’s name. in that recounting. did not come up.
One reason offered for that absence was timing.. By the time ARPANET was taking off, Boehm had left RAND and the field of computer programming.. In 1965. one year after her paper on packet switching was published. she had her first daughter and left RAND to become a stay-at-home mom.. Her husband Barry continued in computer science and was closely involved as ARPANET emerged. while she—according to her daughter—put that chapter away.
Burke described it as a life achievement and a return to priorities shaped by her childhood and her mother being unable to be home as a single mother during the Depression.. Burke also said the death of Sharla’s sister influenced her decision. and that her mother wanted “everyone to be safe and sound.”
For decades, Boehm put her energy into her family and community, especially the Girl Scouts.. Caroline Batzdorf. a friend of Sharla’s daughters. said Boehm was “one of the most understated women” she’d known: someone who quietly encouraged others to strive. supported creativity. and showed up for people without seeking attention.. Batzdorf also said Sharla rarely talked about her computer-programming past; Tenley said she was in her 20s before her mother told her what she had done at RAND.
By the time she had a stroke in 2012, which left her unable to speak, Tenley said she could still hear people and smile. Boehm died in 2023 at age 93.
As the account closes. the contrast between Boehm’s quiet life after RAND and the scale of the work that carried forward is stark.. By the time she died. nearly two-thirds of the world’s population was using the internet that her simulation had helped usher in as a young programmer.. Friends and family. the piece emphasizes. wanted her technology recognized without reducing her to it: she raised a family. nurtured her community. traveled often. and supported girls through their own opportunities.. In Burke’s view. Boehm was proud of her work—and also “quirky” in the way it later turned into the internet. something “bizarre” she didn’t dwell on.
The story keeps returning to the same pressure point: communications that had to keep functioning under extreme damage.. A warning system fails on November 24. 1961. and a relay station’s overheated motor turns a crisis into the need for resilience; the simulation built in the early 1960s tests what happens when nodes are knocked out and whether packets can still find paths; those packet routes then become the technology—packet switching—that ARPANET used after its first incomplete “l-o” message.
Sharla Perrine Boehm RAND Corporation hot potato routing packet switching ARPANET Paul Baran AT&T Strategic Air Command nuclear communications internet origins Lost Women of Science