John Demco: Master of Your Domain

Once upon a time, a loose group of academic techies hooked computers into phone lines to see how far they could reach. They ended up creating the Internet. Vancouver’s John Demco – founder of the dot- ca domain – was one of them.

demco_4.jpg

Once upon a time, a loose group of academic techies hooked computers into phone lines to see how far they could reach. They ended up creating the Internet. Vancouver’s John Demco – founder of the dot- ca domain – was one of them.

The Internet has made itself essential in a very short time. No other medium or technology has ever insinuated itself so deeply into our lives so quickly. Not the telephone, radio, the DVD player nor television. Even the light bulb took longer to catch on – and that was the damned light bulb. And unless you were among the few million people online back in 1993 when the early web browsers came along – odds are you weren’t – the Internet you met when you first encountered it was a relatively easy-to-use point-and-click realm where vast amounts of information were miraculously available and companies or individuals could put up their own websites for a laughably trivial sum of money. Most likely you were too busy exploring it all to think about where the Internet came from or how such a useful thing sprung up fully formed and capable of constantly upgrading itself. But John Demco was there. As com­puter facilities manager at UBC, he helped put together one of Canada’s very first Internet connections. He conceived, petitioned for and secured the dot-ca domain for Canadian websites and administered it until 2000. When the Canadian Internet Registration Authority (CIRA) took over that job, he founded Webnames.ca Inc., which registers and manages domain names for 40,000 clients and employs 20 at its Vancouver office. Now retired from UBC, he serves as director of Webnames.ca. In short, Demco is part of a small group that put Canada on the Web; he’s even got a letter from former prime minister Jean Chrétien that says so. And now UBC is preparing to open a new learning centre named after him – which he’s a bit embarrassed about, he admits. In person, he’s much more humble about his role in creating the dot-ca domain than suggested by his corporate bio on the Webnames.ca website. “I sort of spearheaded the effort, but I don’t want to take too much credit. It was definitely a collaboration,” he says. These days, getting a top-level domain registration like dot-ca recognized is a monstrous bureaucratic ordeal involving world governments. When he did it back in 1987, Demco says, it involved emailing the right guy in California. Jon Postel ran the Internet Assigned Numbers Authority out of the University of Southern California, and his job was to vet and manage a central registry of network addresses. Demco says he found out about Postel through his contacts at networking conferences. “It was much simpler days. There wasn’t even a contract back then. It was very much mostly a trusted relationship.” The funny thing was, the dot-ca domain came to life before Canada had even connected to the major networks forming south of the border that would become the Internet. Demco played a major part in that later on, negotiating with phone companies and the U.S. government to secure a network connection between UBC and the University of Washington in 1986. He successfully made the link-up with the U.S. National Science Foundation’s network two years later. His was among the first such connections in Canada, he says, along with those made by the University of Toronto, McGill University and the Canadian military at around the same time. The very first dot-ca name registry was actually kept on a server in the U.S. because Canada wasn’t sufficiently hooked up to run it. Back then there was no single Internet, he explains; there were a bunch of smaller networks branching out among universities and government agencies, all using their own languages. Demco says he’s still got an old manual lying around from that time, kind of like a phone book with instructions about how to send email to people on different networks. “It was amazingly arcane,” he says. “It was easier for a high-energy physicist at UBC to talk to a high-energy physicist at Stanford or CERN in Europe than it was for him to talk to the computer scientist in the building across campus, because the computer scientist was using another network.” With all these networks in play, getting some consensus on how to use dot-ca was critical, Demco says, “so that I would have some assurance that what I was doing wasn’t going to be so dramatically bad that it was going to cause problems down the road.” The guidelines, technical requirements and application processes were hashed out among the major Canadian computer-network players. Eventually, though, Demco was personally charged with managing all requests for dot-ca domain names, and he did that from 1988 through 2000 (more than 102,000 served, beginning with upei.ca for the University of Prince Edward Island). This effort was a sideline to his regular job at UBC. Was there a sense then that he was part of a momentous innovation that would forever change the world? Demco says yes, but with less drama than you might imagine. The network people knew they were working on something that could be revolutionary, but it happened over a long period of time and there were no guarantees. “I was carefully keeping track of all the [dot-ca] applications,” he says. “I could chart the number of domain registrations per month, or something like that. And if you drew it on a graph, even back in the very early days, you could see that the line on that graph was curving up pretty dramatically.” The whole game changed in the ’90s. The World Wide Web was created and the Internet’s growth was passed from the prudent hands of academics to the invisible hand of the free market. In Canada the major networks were consolidated and fully integrated with those of the U.S. and subsequently those of the rest of the world. “It didn’t happen overnight and so it wasn’t as if there was necessarily an ‘Aha!’ kind of experience. But there was this feeling that, ‘Hey, we are doing something here that actually makes a difference and could actually be a substantial benefit to the country and to people using networks.’ So yeah, that was a really good feeling.” The strength in the development of the Internet, Demco says, “is that problems got solved in a realistic way as they occurred so that things kept on working. With all the [telecommunication companies’] attempts, the tendency had been to try to think of everything you could possibly want to do, ever, and then design the network.” With the Internet, he points out, anything that wasn’t designed at the start created a problem or bottleneck. But as soon as the problem emerged, it would be addressed. “For example, security wasn’t an issue in the beginning. They were much more concerned with things like reliability in difficult conditions. They did deal with a lot of the really important issues, but how can you plan something indefinitely into the future?” Compared to the so-called generic domains, such as the well-known dot-com, that are open to anyone, it used to be tough to get a dot-ca. “A lot of the country code domains started off with a very restrictive process,” Demco claims, citing presence requirements (you had to have offices in two or more provinces) and limits on the number of domains you could own. “One was a very common restriction in many countries, but we had a presumed limit of two, to accommodate French and English. CIRA came about when it was recognized that the little volunteer effort I was heading wasn’t going to continue to scale up and also recognizing that we probably needed to change the rules anyway. They were for another era.” Still, Demco thinks it was an advantage for Canada to be a bit late in loosening its regulations, because it helped protect the integrity of dot-ca as a brand. He points out how the dot-net and dot-org domains – which used to be reserved for qualified groups such as Internet providers and non-commercial entities – lost their meaning when they were opened up in the dot-com rush. By and large, dot-ca has a positive connotation online and perhaps also serves as a kind of aptly low-key form of Canadian nationalism among those who choose to use it. One million domains with the dot-ca name are expected to be registered by the end of the year or so, which lends a bit of exclusivity compared to the 50 million dot-com names in circulation. That leads to one of the Internet’s major weaknesses: there’s a limit to the number of Internet Protocol (IP) addresses available under the current system, and, with everyone and their kitchen appliances threatening to need a place online, “the prediction is that, at the current rate of allocation, we’ll run out of address space in three or four years,” says Demco. That’s different than the domain-name problem (like names for rock bands, all of the good ones are already taken). IP addresses are like phone numbers; each one is attached to one specific place. Machines called domain-name servers look up that number when you type a name into your browser or send some email. There is a new standard ready to go (equivalent to adding two more area-code slots), but upgrading now “is somewhat akin to changing the engine of a plane in mid-flight.” However, that upgrade day needs to arrive (it’ll be Y2K-big in advance and ideally Y2K-small in retrospect), Demco warns, or IP addresses will become a scarce and possibly expensive commodity when demand outstrips supply, restricting access and the Net’s usefulness. Assuming that problem is resolved, Demco feels the next major issue will be the need to vastly increase the speed with which ordinary users can access the Internet from their homes, enabling high-bandwidth communications, telepresence – the concept of making remote things and experiences feel close, collaborative projects and really big video display walls (he has a bit of an obsession with those). “The bottleneck that will become apparent over the next while is the last mile, the connection between your home and the next point up the line. That hasn’t changed in years,” Demco complains. “It’s still either twisted-pair [wires] from the telcos or a tiny little slice of the cable infrastructure. It’s a serious limitation that, if it’s not already, will very soon be holding up the kind of innovation that we’ll want to be able to do.” But making predictions about the Internet can be tricky, he says. (This coming from the guy with the foresight to secure dot-ca before Canada was even connected to what would become the Internet.) “I’m certainly interested in trying to figure out what are going to be some of the next big steps. And sometimes this is impossible – because who can predict the future?” – with files from Peter Severinson