'We were just trying to get it to work': The failure that started the internet

 On  October 29, 1969, two scientists connected computers  350 miles apart and started typing messages. Halfway through, the devices crashed.  55 years later, they spoke to the BBC. At the height of the Cold War,


Charlie Klein and Bill Duvall were two clever engineers at the forefront of technology's most ambitious experiments. Klein, a 21-year-old graduate student at the University of California, Los Angeles (UCLA), and Duvall, 29, a  systems programmer at the Stanford Research Institute (SRI), worked on a system called Arpanet for the Advanced Research Projects Institute Network. Funded by the U.S. Department of Defense, the project aimed to build a network that could exchange data directly  without using telephone lines. Instead, the system used a data transfer method  called "packet switching," which later formed the basis of the modern Internet. It was the first test of a technology that would change nearly every aspect of human life. But before it worked, you had to log in. Klein sat at his keyboard between the lime-green walls of Room 3420 in UCLA's Bolter Hall, preparing to contact Duvall, who worked with computers across  California. But Klein hadn't even finished saying the words "L-O-G-I-N" when Duvall called to tell him the  system had crashed. Thanks to this mistake, the first "message" Klein sent to Duvall on that fall day in 1969 consisted solely of the letters "L-O." After some tweaks, about an hour later the connection was back up and running. The initial crash was just a blip in an otherwise monumental achievement. But neither man realized the significance of the moment. "We certainly didn't realize it at the time," Klein says. "We were just trying to make it  work." The BBC interviewed Klein and Duvall to mark the 55th anniversary of the event. Half a century later, the Internet has shrunk the whole world into a tiny black box that fits in our pocket, dominating our attention and touching the furthest corners of our life experience. But it all started with two men experiencing the frustration of not being able to get online for the  first time. This interview has been condensed for clarity and length. Can you describe the computer that made the Arpanet possible? Was it a huge, noisy machine? Klein: It was a tiny computer for its time, about the size of a refrigerator. The cooling fans made a little noise, but it was quiet compared to the fan noise in a Sigma 7 computer. On the front it had flashing lights, a switch to control the IMP [Interface Message Processor], and a  tape reader for loading software.

Duvall: It was housed in a rack big enough to house a full sound system for a big show today. And they were thousands, if not millions, billions, of times less powerful than the processor in the Apple Watch. Those were the  days! Tell us about the moment  you started typing L-O. Klein: Unlike today's websites and other systems, when you plugged a terminal into the SRI system, nothing happened until you typed something into it. If you wanted to run a program, you first had to log in by typing the word "login," and the system would prompt you for your username and password.  I typed characters into the terminal (a Model 33 teleprinter), and the terminal sent the characters to the program I'd written for the SDS Sigma 7 computer. This program takes characters, formats them into a message, and sends it to the interface message processor. When the SRI system receives it, it treats it as if it came from a local terminal and processes it. It echoes the characters [duplicates them on the terminal]. In this case, Bill's code takes the characters, formats them into a message, sends it to the IMP to send back to UCLA, and then receives it and prints it on the terminal.  I was on the phone with Bill when I tried this. I told him I typed in the letter "L". He told me he  received the letter L and sent it back. I told him I printed it. Then I typed in the letter "O". Again, it worked perfectly. I typed in the letter G. Bill said the system had crashed and he'd call me back. Duvall: The UCLA system wasn't expecting to receive G-I-N after Charlie had gone L-O, so it sent an error message to the SRI computer. I don't remember exactly what the message was, but what happened next was  that the network connection was so much faster than anything I'd ever seen before.  The normal connection speed was 10 characters per second, but the Arpanet could send up to 5,000 characters per second. When this message was sent from UCLA to the SRI computer, the input buffer filled up with only the expected 10 characters per second. It was like filling a glass with a fire hose. We quickly realized what had happened, changed the buffer size and rebuilt the system. This took about an hour. Klein: No, we certainly didn't do that back then. Duvall: Not really. This was a new step  in the larger context of our work on SRI, and I was convinced it would have a big impact. When Samuel Morse sent the first telegraph message in 1844, he focused on the dramatic, typing "What  God Has Done" onto the wire from Washington, D.C.

to Baltimore, Maryland, USA. If you could go back, could you have typed something more memorable? Klein: Of course, if I had known the meaning. But we were just trying to make it  work.Duvall: No. This was the first test of a very complicated system with many moving parts. The fact that something so complicated worked on the  first test was dramatic in  itself. What was the mood like when you sent the message? Duvall: We were both alone in our respective computer labs at night. We were both pleased that after a lot of work, the first test was successful. I went to the local "pub" and had a burger and a beer. Klein: It worked, so I  went home and got some sleep. What did you expect from Arpanet? Duvall: I saw the work we were doing at SRI as an important part of a larger vision of information workers connecting and sharing problems, observations, documents, and solutions. What we didn't see coming was its commercial acceptance, nor did we anticipate the  social media phenomenon and the proliferation of disinformation  that came with it. However, [SRI computer scientist] Douglas Engelbart wrote in a 1962 paper outlining the overall vision: The capabilities we are creating will cause profound changes in  society, and what is needed is for the tools we create to be adapted to address the problems that arise from using them in society.  These were thousands, if not millions, billions, of times less powerful than the processors in your Apple Watch. Those were the  days! -Bill Duvall In what ways does today's Internet remind you of the Arpanet? Duvall: In terms of the larger vision developed in Engelbart's group (the mouse, full-screen editing, links, etc.), today's Internet is the  logical evolution of those ideas, expanded of course by the contributions of many smart, innovative people and organizations. Klein: The ability to leverage resources from others. This is what we do when we use a website. We use the website and its programs, features such as functions, and of course email.  Arpanet basically created the concept of routing and multiple paths from one site to another, which ensured reliability in the event of a communication line failure. It also made it possible to increase communication speeds by using multiple routes simultaneously. These concepts were transferred to the Internet. In developing the communication protocols for the Arpanet, we discovered problems, corrected and improved the protocols, and learned a lot of things that we were able to transfer to the Internet. TCP/IP [the basic standard for Internet connections] was designed to connect networks, especially the Arpanet, to other networks, improving performance, reliability, etc. What do you think about this anniversary? Klein: It's a mix. Personally I think it's important, but a little bit excessive. The Arpanet and what came out of it is very important. For me, this anniversary  is just one of many events. What I think is somewhat more important than this  anniversary is Arpa's decision to continue building the network and supporting its development.  Duvall: It's good to remember the origins of something like the Internet, but what's most important  is the tremendous effort that's been made since then to make the Internet an important part of society around the world.  The modern web is controlled not by governments or academic researchers, but by  the world's largest corporations. How do you feel about what the Internet has become? What worries you most?Klein: We use it in our daily lives and it's very important. It's hard to imagine  not having it anymore. One of the benefits of  being so open and free of government control is that it allows new ideas to develop: online shopping, online banking, video streaming, news sites, social media, and more. But because it has become so important to our lives, it has become a target for malicious activity.  You hear all the time about how things are being compromised. There's a huge loss of privacy. And the big corporations (Google, Meta, Amazon, and internet service providers like Comcast and AT&T) have too much power in my opinion. But I don't know what the right antidote is.  Duvall: I think  there are great dangers in the dominance of any single organization. We've seen the power of disinformation in shaping politics and elections. We've also seen the power corporations have in shaping social norms and the development of adults and young people. Klein: One of my biggest fears was the spread of misinformation. How many times have you heard someone say, "I saw it on the Internet?" It's always possible to spread false information, but it costs money to send out advertising letters, put up billboards, or run TV ads. Today it's cheap and easy. And when that information reaches millions of people, it gets repeated and treated as fact.  Another concern is that as more and more critical systems move online, they become more susceptible to causing significant disruptions if they fail or are  compromised. This applies not only to communications systems, but also to banks, utilities, transportation, and more.  Duvall: The Internet has great power, but we have ignored Engelbart's 1962 warning and have not effectively harnessed the power of the Internet to address societal impacts.  Are there any lessons you've learned from your time at Arpanet that can make it a better place for everyone?Klein: The Internet is open, which allows for experimentation and new uses, but also opens the door to compromise if not controlled. Arpa maintained some control over Arpanet: this way they could make sure  everything was working, make decisions about required protocols, and address site names and other issues.

Learn more  Google just updated their algorithm. The Internet will never be the same again "Postcards are the email of their day": How cat memes spread 100 years ago  We're losing our digital history. Can the Internet Archive preserve it?  Icann (Internet Corporation for Assigned Names and Numbers) still manages some of it, but there's international disagreement about how to proceed and whether the U.S. has too much control. But some control is still needed to keep the network functioning. Additionally, because the Arpanet was relatively small, it was able to experiment with big changes in design, protocols, etc. that would be very difficult to do today. Duvall: AI and its reflexive access to everyone who uses the Internet are what are pushing us to the edge of  the abyss. The Internet experienced explosive growth and development in its early days, some of which was societally damaging. AI has now reached this threshold and is inextricably linked to the Internet. And it is not unreasonable to describe AI as an existential threat. Now is the time to recognize the threat and the opportunity.

Post a Comment

Previous Post Next Post