Expect to see some unfamiliar syllables after the dot on the Internet this spring. Besides the .com, .net, .org, .edu, .mil and .gov the world has come to know, seven newcomers are about to enter the fray: .aero, .biz, .coop, .info, .museum, .name and .pro. These additions to the list of so-called top-level domains were approved last November by the Internet Corporation for Assigned Names and Numbers, or ICANN-the closest thing the Internet has to a government. This organization’s stated purpose in creating fresh cyber real estate was to open opportunities on the Internet for business and noncommercial use alike. But despite the good intentions behind them, the new dots are likely to remain a footnote in the Internet’s evolution.Domain names were not part of the original Internet design. They were, rather, an afterthought-created as a techno-fix to a pressing problem by engineers who didn’t foresee the long-term implications of their actions. But now, domain names are vital because Web browsers use them as the basis of all navigation. Want to jump to the home page for the White House? Just type the domain name “whitehouse.gov” and off you go. Domain names are the coin of the cyber realm; stories abound of companies trying to buy their domains from so-called cyber squatters, and of large organizations trying to shut down small Web sites because of similar-sounding names.
But the Internet’s underlying structure doesn’t really use long-winded names to move data about; it uses a compact numbering system. “Whitehouse.gov,” for instance, is a human-friendly representation of the address 22.214.171.124. The job of translating the domain names that people type into their browsers into the Internet addresses that the network actually uses falls to an entity called, logically enough, the Domain Name System, or DNS.
The original Internet had no such system. Instead, the addresses for every computer on the Net were kept in a single “hosts” file on a computer at the Stanford Research Institute in northern California. If you had a computer on the Internet, and wanted to maintain an up-to-date version of what else was out there, it was your responsibility to download this file regularly. You can think of the hosts file as the Internet’s first white pages.
By 1982, the Internet was growing so rapidly that nobody had a current copy of the hosts file. But there was another problem as well: sometimes more than one computer in the file had the same name. A friend of mine named Martha Rose, for instance, had the mail account “mrose” on the computer “Eddie” at MIT; Martha was forever receiving e-mail for Marshall Rose, who had the same account on a computer named Eddie at the University of Washington. Both computers were in the hosts file under the same name.
The Domain Name System was supposed to solve both of these problems because it is not a single file, but rather a series of files in a database distributed across many computers on the Net. Computers at MIT called name servers hold the Internet directory for MIT; similar servers at the University of Washington hold its directory. The distributed database worked spectacularly well-so well, in fact, that it has not been significantly upgraded in more than 15 years.
The system was also supposed to put an end to naming conflicts of the sort that vexed my friend Martha. The Internet’s engineers created the .edu domain for universities, under which each university would have its own unique sub-domain. Martha’s e-mail address became email@example.com; Marshall’s became firstname.lastname@example.org. Likewise, the engineers created the top-level domains .com, .gov, .mil, .net and .org. Because of the U.S.-centric nature of these domains, the system soon expanded to include more than 200 additional top-level domains, one for each country.