Summary

  • The Cotonou summit gave Benin’s Internet connection a deadline, but sources differ on whether the first connection came in November or December 1995.
  • OPT’s gateway preceded routine service and ISP growth; later providers could multiply while still depending on OPT’s international circuits.

Mail could arrive before the country had its gateway

“Benin connected to the Internet in 1995” sounds like a clean starting line. It hides several different thresholds. Ken Lohento’s retrospective chronology says a small number of Beninois had exchanged email through BOSS using UUCP from 1994. That was meaningful communication, but store-and-forward email over UUCP was not the same thing as a routed, continuously reachable national Internet gateway. The distinction matters: 1995 did not mark the first electronic message, and a gateway did not suddenly give every person an online connection.

The useful question is narrower: when did an international Internet path begin operating, who operated it, and what did that arrangement make possible afterward? On those questions, the record points to the Office des Postes et Télécommunications, usually called OPT, and to a connection prepared around a diplomatic event.

The summit put a date on infrastructure

Cotonou hosted the sixth Francophonie summit in 1995. A historical survey by Annie Lenoble-Bart and André-Jean Tudesq dates OPT’s first connection to November, during the summit period. Lohento’s later account places the gateway in service for the summit in December. A UN-hosted national report also gives December. The difference is not enough evidence to choose an exact commissioning day; “late 1995” is the safer description.

The summit supplies a plausible operational deadline, not proof that diplomacy alone caused the connection. A major gathering brings visiting officials, journalists and organizers who need international communications, creating a reason to have a path ready at a particular moment. Lohento describes a connection through France Câbles et Radio, but sources disagree about initial capacity: his account gives 64 kbit/s, while the UN report states 1 Mbit/s. Those may refer to different points or stages, yet the record does not reconcile them. The article therefore treats the capacity as unresolved rather than choosing the more dramatic number.

The event connection and the public service that followed were also different milestones. Lohento and Théophile Vittin’s 1999 account both place OPT’s regular Internet service in February 1996. That interval matters. A link can be installed or used for an event before an organization has settled service, billing, support and access arrangements for ordinary customers. “Connected” names a technical condition; “available as a service” names an institutional one.

The gateway outlasted the event

As providers and access points appeared, OPT’s role did not vanish. A 2001 response preserved in the NSRC Connectivity Providers Database says OPT established the external gateway in 1995 and remained the principal gateway, with private ISPs using OPT-linked circuits. It is a later snapshot, not a full diagram of the 1995 network. Still, it shows why counting providers alone can mislead: a market can acquire more retail brands without acquiring multiple independent paths to the global Internet.

Vittin’s 1999 account describes growth in subscribers, providers and cybercentres, including service beyond Cotonou. Those figures should be read as that report’s administrative estimates, not a census of people online. They conflict with an ITU country report that lists 1,537 subscribers and four ISPs at the end of May 1999. The dates differ, and the sources do not supply definitions that would make the totals comparable. Shared cybercentre use also means accounts cannot be equated with users. The evidence supports expansion; it does not support a single reconciled national penetration figure.

The centrality of OPT should likewise be stated precisely. The evidence describes an upstream gateway and dependence on OPT circuits. It does not prove that OPT ran every provider, local access point or user connection, or that no alternative technical route existed at any time. Nor does it show how often the gateway failed or what users paid. A national route can be central without being the whole network.

The longer history is therefore not simply “summit, then Internet.” It is a sequence: limited UUCP mail; a late-1995 gateway tied to an event deadline; routine OPT service in early 1996; expanding providers and shared access; and a wholesale arrangement that remained concentrated. Using Heng Lu’s note on running systems as an editorial lens, rather than as evidence, clarifies the distinctions: an announced connection, an operating gateway, a routine service and a person’s usable access are separate facts. Cotonou brought the first of these Internet thresholds into view. OPT’s continuing gateway role shaped what came next.

Sources