Summary

  • A participant-maintained HEAnet chronology dates a 9.6-kbps EUnet Internet line at Trinity College Dublin to June 1991, running with TCD and reaching the Internet through Canterbury and EUnet in Amsterdam. It says TCD and Ieunet were directly connected, while UCD joined through IXI around the same time; the chronology only estimates that TCD was ahead by weeks.
  • Niall O’Reilly recalls UCD’s HEAnet-2 as a multi-protocol X.25 network that encapsulated IP and IPX. UCD’s two IXI links served different communities, HEAnet sites and EuroKom, and could carry encapsulated IP to NIKHEF in Amsterdam. The TCD archive preserves the route chronology. O’Reilly’s account explains UCD’s X.25 path.

The distinction is easy to lose in a national timeline: “connected” described more than one engineering arrangement. The HEAnet Dateline records that Trinity College Dublin and Ieunet had a 9.6-kbps line to the Internet in June 1991. EUnet operated it with TCD; traffic reached the wider Internet via Canterbury (the University of Kent) and EUnet in Amsterdam. The same chronology describes TCD and Ieunet as directly connected, and UCD as connected through IXI around that time. Its compiler thought the TCD route preceded UCD’s by several weeks, not by a securely established interval.

UCD’s route was different in kind. Niall O’Reilly says he returned to UCD in July 1991 and began work on HEAnet-2, designed as a multi-protocol network over X.25, with IP and IPX encapsulated above it. He recalls two IXI connections at UCD: one for HEAnet sites and another for EuroKom. Those links also provided Internet access through an X.25-encapsulated IP connection to the IXI access point at NIKHEF in Amsterdam. The recollection explains the path, but does not pin down the exact date when UCD’s service began.

The archived chronology places this design amid a deliberate change in HEAnet’s scope. In December 1990, HEAnet adopted a policy of connecting campus LANs rather than just individual machines. The following March, its Futures committee proposed router-based HEAnet-2 carrying several protocols, mostly TCP/IP, over X.25. This was a design proposal and deployment sequence, not evidence that every campus LAN immediately had the same access. The chronology also describes universities using 64-kbps Eirpac VPNs in 1991 and says the HEAnet agreement was not volume-charged, though it does not supply usage records.

John Boland’s account from Dublin City University shows why the packet network mattered to applications. He recalls early-1990s SMTP, FTP and Telnet traffic moving over a 64-kbps VPN on Telecom Éireann’s Eirpac packet-switched network. Telnet’s small packets and repeated handshakes could affect Eirpac unpredictably; the operator eventually wanted the traffic moved. This is an operator’s recollection, not a dated incident log, and it should not be projected onto every HEAnet route. Boland describes the Eirpac constraint.

The lesson is not that X.25 prevented the Internet from working. It carried traffic and provided an international path where a dedicated line was not the only option. But reachability depended on the transport layer, the gateway and the institution using it. TCD’s EUnet line and UCD’s IXI route did not amount to one national switch being turned on; later Eirpac experience shows that a path could function while application behavior still strained it. HEAnet’s own institutional timeline places its formation in 1983 and UCD’s management contract in 1993–97, useful context for the network’s collaborative roots rather than detailed proof of 1991 routing. HEAnet’s institutional history gives that broader chronology.

Heng Lu Note 65 is an editorial lens here: distinguish a proposed design, a participant’s recollection of a running route and an observed application problem. It is not evidence about Ireland. The available sources do not settle a single national “first Internet” date, prove household availability or independently measure how much traffic used either 1991 path.

Sources