Summary

  • In 1995, a Fido node at CNAR in N’Djamena exchanged mail through a Paris relay; this was store-and-forward messaging, not evidence that Chad already had a public Internet gateway.
  • A UK traceroute reached bow.intnet.td on 4 November 1997, while SOTEL Tchad later dated Tchadnet’s opening to 19 November. A single host response establishes a path, not nationwide service.

The chronology looks contradictory only if “connected” is treated as one switch. The records describe different technical states. A contemporary note sent to the Network Startup Resource Center in November 1995 says a Fido node had been installed at the Centre National d’Appui à la Recherche (CNAR) in N’Djamena. Its address was 5:7791/1; a subdomain had been requested through GreenNet, and mail travelled through Globenet, a Fido node in Paris. The correspondent said direct international calls were then possible only to France and Cameroon in their circumstances, with poor quality to Cameroon and a good transfer rate when a call to France succeeded. That is a valuable account of one operator’s working conditions, not a survey of every institution or household.

Fido mail could be composed locally, queued, and relayed when a telephone connection became available. It did not require each correspondent to be online simultaneously, and its Paris hop did not make the N’Djamena node an Internet gateway. A domain-like address and an international relay are evidence of organized communication, but neither alone establishes routable IP service to the public.

The next record captures a different threshold. An NSRC report dated 4 November 1997 includes a traceroute from a UK host to bow.intnet.td (194.250.40.253). The reported path crossed London academic-network links, a Telehouse/LINX segment, and Paris-area hops associated with France Télécom before the target answered. Final round-trip readings were roughly 0.9 to 1.06 seconds. This is unusually concrete evidence: a named Chad host was reachable from a particular origin at a particular time. It cannot tell us how many other hosts were reachable, who inside Chad could use them, or whether domestic routes were robust.

SOTEL Tchad’s presentation to an ITU workshop in 2004 retrospectively places Tchadnet’s opening on 19 November 1997, with a connection to France Télécom’s Bagnolet node in Paris. It reports a 64-kbit/s full-duplex gateway and 43 subscribers at the end of that year. Thus, the traceroute predates the operator’s recorded opening by fifteen days. The surviving sources do not explain the exact status of bow.intnet.td during that interval. It may have been part of setup or testing, but that remains an inference, not a documented explanation.

The distinctions matter. A Fido node exchanging mail, a host answering an IP trace, a national gateway’s formal opening, and a counted subscriber base are separate measures. The 64-kbit/s figure is aggregate full-duplex capacity; dividing it by 43 would not yield observed user throughput. Nor should SOTEL’s 2004 description of satellite dependence and heavy concentration of telephone infrastructure in N’Djamena be projected backward as a measured 1997 topology. The presentation lists later provincial nodes as work or project scope, not proof that all were operating.

The careful conclusion is modest but significant: before Tchadnet’s recorded opening, Chad already had a reported mail relay and at least one demonstrably reachable host. The gateway then gave the country a documented international Internet connection and an initial subscriber count. These traces mark growing operational capacity; they do not establish broad or geographically even access.

Sources

The Heng Lu note informs the interpretive distinction between a running route and an institutional plan; it is not evidence about Chad.