Summary

  • The April 1993 connection gave Turkey a documented path to NSFNET; it did not turn one leased line into a household network.
  • The later TURNET contract, gateway rules and academic handoff mattered as much as the line’s headline capacity.

A line is an event, not a whole network

The chronology starts before the celebrated 1993 connection. ODTÜ and TÜBİTAK began the TR-NET project in 1991. The METU Internet archive records an X.25/IP application to the Netherlands in 1992 and dates the first X.25 Internet test to October. A later TÜBİTAK retrospective gives September for that test. The sources agree on the sequence, not the month; the discrepancy should remain visible rather than be made to disappear. (METU archive; TÜBİTAK retrospective)

The next milestone is more precise. METU records that PTT supplied the 64 kbps leased line on 5 April 1993 and that the ODTÜ-to-NSFNET TCP/IP connection opened on 12 April. METU’s Computer Center describes it as a link that served the country, but that institutional phrase does not document household coverage, retail availability or a last mile to every user. The capacity describes one international connection, not the reach of every domestic institution. (METU Computer Center history)

Nor did users wait for TURNET. Wolcott and Çağıltay’s 2001 study reports that, by early 1995, TR-NET had about 3,000 hosts, 10,000–15,000 daily users, more than 100 connected institutions and over 1,300 individual connections. Those are the authors’ historical estimates, not an audited census, but they rule out the neat story in which a commercial backbone created Turkey’s first Internet users. TR-NET had already grown through universities and other organizations. (Wolcott and Çağıltay, 2001; Yaşar Tonta, 1996)

From a proposed backbone to a contract

A 1995 TR-NET proposal sketched a stronger domestic backbone, including a proposed Ankara–Istanbul–İzmir triangle and a service-provider layer. A proposal is evidence of intended design, not proof that those links were built. The commercial path took shape through a separate Türk Telekom tender: Wolcott and Çağıltay date its announcement to 28 September 1995, the auction to October–November, and the award to a consortium of GlobalOne, Satko and METU. The seven-year contract was signed on 1 March 1996. The paper reports a 70.2% initial revenue share for Türk Telekom, rising to 79.6% in year seven. These figures describe the reported contract arrangement; they should not be mistaken for a complete account of who funded or owned every asset. (Wolcott and Çağıltay, 2001; Çağıltay’s participant retrospective)

The study describes TURNET’s operating design as three 2 Mbps domestic links and two 512 kbps international connections, from Istanbul and Ankara. Internet service providers had to connect through TURNET and send international traffic through its gateways; independent satellite links were not permitted. That arrangement lowered the cost of entry for an ISP that no longer needed to build its own international connection. It also concentrated international access in one shared backbone and its carrier-controlled gateways. Both effects belong in the same account.

The date of operation is less tidy than the contract. TÜBİTAK’s retrospective says the TURNET project was launched in August 1996. Tonta’s contemporaneous 1996 paper says service began in September. Wolcott and Çağıltay call it a fall 1996 start. These accounts may describe different steps, but the sources do not establish a common operational date. “Fall 1996” is the defensible summary; a precise day is not.

The academic path did not vanish

A second handoff was under way. ULAKBİM was established in May 1996 and began operating on 1 June, with an academic-network role. Its later ULAKNET backbone is a separate milestone: TÜBİTAK ULAKBİM dates the 34 Mbps ATM backbone’s launch to 1997. The commercial TURNET and the academic ULAKNET should not be collapsed into one successor network. They served different institutional purposes and reached operational milestones at different times. (TÜBİTAK ULAKBİM history; Yaşar Tonta, 1996)

The importance of this episode lies in the distinctions it preserves: a test connection, a leased international line, a growing research network, a tender, a signed revenue-sharing contract, a commercial backbone and an academic successor. The 64 kbps figure is a useful historical marker, but it cannot stand in for all the institutions and interconnection rules that made connectivity usable. A 2018 METU oral-history catalogue confirms that participants recorded the period from the project proposal through the first connection; it is evidence that an oral-history project exists, not a substitute for a published transcript. (METU oral-history catalogue)

Sources