Summary

  • NIHERST reported a dial-up UUCP connection to UPR2 in Puerto Rico for email and FTP, with its divisions and the Institute of Marine Affairs (IMA) reaching the outside through NIHERST’s head-office node.
  • The institute said IMA’s local telephone costs were recovered without profit according to mail volume; the record does not state the total bill, link speed or message count.

A network can be international at one end and intensely local at the other. In a 31 December 1996 update archived by the Network Startup Resource Center (NSRC), Trinidad and Tobago’s National Institute of Higher Education, Research, Science and Technology (NIHERST) described a dial-up UUCP connection to UPR2 in Puerto Rico. It used the link for email exchange and FTP file downloads. The same report shows that the route’s institutional boundary did not end at the modem that called abroad: NIHERST’s divisions and an outside research institute reached it through a node at NIHERST’s head office. [S1]

That arrangement matters because the Institute of Marine Affairs (IMA) did not initially have International Direct Dialing at its location. NIHERST said IMA therefore connected to the Internet through NIHERST and UPR2. Its departments—the Chemistry Department, Document Centre, Office of the Vice President and another division identified as SOL—dialed the head-office node to send and retrieve mail. The archive does not explain the SOL abbreviation or the exact local dial plan, so the report supports a shared gateway, not a detailed map of every call. [S1]

The accounting rule is unusually concrete. NIHERST said it recovered costs from IMA “with no profit,” based entirely on the volume of mail sent to and received from that node. The money covered local telephone charges for uploading and downloading the messages to UPR2. This is a volume-based allocation rule for phone costs. It is not evidence of a per-byte Internet tariff, a metered wholesale service, or a complete cost model: the report supplies no message totals, call duration, local tariff or monthly amount. [S1]

The equipment list hints at a distributed, modestly provisioned system without establishing its throughput. NIHERST listed a 486-class head-office machine with a 14.4 modem, a 386 with a 14.4 modem at its Chemistry Department, a 286 with a 14.4 modem at the Office of the Vice President, a 486 and 9600 modem at its Document Centre, and a 486 with a 28.8 modem at IMA. These are endpoint descriptions. The source does not say which modem negotiated each connection, whether the international leg used the same hardware, or how much mail moved. No international UUCP speed should be inferred from the list. [S1]

A separate NIHERST update from 23 December 1996 placed the arrangement in a wider, but still respondent-reported, picture. It named several providers using TSTT services and said many academic and research institutions had no Internet connection, or only a simple dial-up link to a provider. The correspondent believed those institutions did not then need a full leased line. That statement is not a census and does not negate NIHERST’s more specific UUCP route: “connection” covered more than one service model, while institutional demand and external reach could differ. [S2]

A year earlier, a UK-origin trace reached TSTT’s name server through SprintLink and recorded 25 percent packet loss in that particular ping sample. The trace demonstrates reachability of one public-facing host from one location on 1 September 1995. It says nothing about NIHERST’s later email circuit, service quality elsewhere in the country, or whether a research office had an interactive connection. The difference between the records is itself useful: a reachable provider host and an institute’s scheduled store-and-forward email were different observations, not interchangeable measures of “Internet access.” [S3]

The surviving evidence therefore supports a narrow chronology: by late 1996, NIHERST reported an email-and-file route to Puerto Rico, local offices and IMA used its head-office node, and the institute allocated IMA’s local telephone costs by message volume. It does not identify when the arrangement began, how often the calls ran, what it cost in total, or when IMA obtained direct dialing. The evidence is a contemporaneous institutional response preserved in an archive, not an audit or a national usage survey.

Sources: NSRC Trinidad and Tobago archive index; NIHERST Connectivity Update, 31 December 1996; NIHERST/Trinidad and Tobago networking update, 23 December 1996; UK traceroute and ping to TSTT, 1 September 1995; New IP hosts TT, 1 September 1995; Heng Lu Note 65 (editorial lens only).