Summary

  • Cleveland Free-Net grew from a 1984 medical question-and-answer experiment into a no-fee community computer system; CWRU dates its first Internet access to the 1989 rebuild, not its 1986 opening.
  • In 1995, CWRU estimated a telephone-and-modem session at about five cents in combined system and telephone costs, while an Internet-routed session cost the system less than one-tenth of a cent. The system’s modems were already busy for more than 99% of available time.

Analysis

In 1984, Thomas Grundner and colleagues at Case Western Reserve University began with a single-line medical bulletin board. A caller could leave a health question for a physician to answer. Two years later, the experiment had become Cleveland Free-Net: a public computer system that offered email and information on subjects including health, education, government, arts and law. The university’s historical account dates the opening to July 1986. It describes a community service people could reach with a computer or terminal and a modem, years before the public Web.

“Free” described the price of using the service, not the absence of an access chain. A household needed a computer or terminal, a modem and a telephone connection; the host needed hardware, phone lines, software and people who kept the system useful. A 1989 research report on community networks describes Cleveland’s prototype with more than 6,000 registered users, 500 to 600 logons a day and ten incoming telephone lines. Those are three different measures: accumulated registrations, daily logons and simultaneous paths into one host. None alone tells us how many residents could get through at a busy hour.

Growth forced a change in the machine behind the public promise. Case Western Reserve’s institutional history says a capacity reorganisation was completed in August 1989 and added Internet access. The system therefore crossed a boundary: a local community computer service that had opened in 1986 gained a route to the wider Internet three years later. That change did not make the first version an Internet service retroactively, or turn every internal discussion and information service into general Internet access.

By June 1995, CWRU’s Office of Information Services was reporting a much larger service. Its comparison of 1988 with 1994 put individual users on a given day at 160 and then 16,000, and annual sessions at 36,000 and then nearly six million; the average session lasted 28 minutes. The same account said the system could support 406 simultaneous users across access methods and had 260 dial-in telephone lines. These figures describe different things, so the registered-user count, daily user count, session count and concurrent capacity should not be collapsed into one measure of “users.”

The access route changed the economics. CWRU’s 1995 report said more than half of all use was then arriving through the Internet. For a session reached through a telephone and modem, it estimated less than one cent in system cost plus about four cents associated with use of the public telephone system: roughly five cents in total. An Internet-routed session cost the system less than one-tenth of a cent. This was not a bill sent to every user. It was an operating estimate that made visible who absorbed costs the no-fee login screen did not show.

The lower-cost route did not end the capacity problem. CWRU reported that its modems were in use for more than 99% of all available time and said it could use twice or even three times as many. The university had spent more than $150,000 on computer hardware across three generations of upgrades; licensing its FreePort software to other communities had raised about $60,000, all of which it said went back into hardware. Users also donated used modems, but the university eventually had to buy hundreds of replacements. Free access depended on continuing capital, donated equipment and institutional support.

That pressure helps explain the ending without making it a simple story of the Web replacing a quaint bulletin board. CWRU’s historical account says that heavy demand brought malfunction and frequent busy signals even after upgrades, while commercial Internet access was becoming more affordable. The university closed the original system on 30 September 1999 after identifying the cost of further performance improvements and Y2K compatibility. The shutdown notice told community users to find other ways to receive email and use Internet resources.

The model travelled beyond Cleveland. A contemporaneous history from Ottawa’s National Capital FreeNet says its organisers acquired software from CWRU in 1992 and opened their system publicly in February 1993. That is evidence of a transferable design, not proof that every later Free-Net had the same costs or operating conditions.

The useful historical unit is the session and the route that carried it. A nominally free public service still had to pay for the computer, telephone interface, support and renewal. Cleveland’s Internet route cut the reported cost per session by orders of magnitude, but the system still faced near-saturated modem capacity and a funding limit. “Free” was a deliberate allocation of cost. It was not a technical property of the network.

Sources