Summary

  • ERNET began in 1986 as a cooperative project among eight Indian academic and government institutions. Its early mail service, international gateways and campus links grew in stages; they did not give every Indian user interactive Internet access at once.
  • The network’s institutional reach preceded the public service launched through VSNL on 15 August 1995. That date marks a new access offer, not India’s first connection to the Internet.
  • The consequential design choice was who could participate, operate links and sustain them—not simply when a line was switched on.

A national network began with a defined circle

In 1983, computer scientist Srinivasan Ramani proposed a nationwide academic network. The Department of Electronics took up the idea, and in 1986 the Education and Research Network project—ERNET—began with Indian government and United Nations Development Programme support. Its eight participating institutions were the National Centre for Software Technology in Bombay, the Indian Institute of Science in Bangalore, five Indian Institutes of Technology, and the Department of Electronics in New Delhi. That list defined a starting coalition, not a promise that every university or researcher already had a terminal and a usable connection.

Ramani’s retrospective describes a project in which the institutional coordinators worked as collaborators, with young engineers and students carrying much of the learning and implementation.

The earliest service makes the distinction concrete. ERNET began with dial-up and UUCP mail exchange. A message could be prepared locally and passed between systems when a scheduled connection was made. That was valuable international communication, but it was not the same experience as a user keeping an interactive IP session open to a remote host. The 1998 account by Grey Burkhart, Seymour Goodman, Arun Mehta and Larry Press likewise distinguishes India’s email users without full TCP/IP connectivity from people with full IP access.

Any history that compresses these services into a single “Internet arrived” date erases the operating conditions users actually encountered.

Gateways grew by stages

At the international edge, the system evolved rather than appearing fully formed. The Internet Hall of Fame’s biography of Ramani places ERNET’s first international link in 1987: TCP carried over X.25 to Amsterdam. In 1988, Ramani’s team connected to UUNET in Falls Church, Virginia. Partner institutions then worked with the NCST team to establish domestic TCP/IP links. These were different steps: a mail exchange, a gateway that could speak TCP over an existing packet network, and links that extended IP connectivity among institutions should not be treated as synonyms.

A March 1995 message from Ramani, preserved by the Network Startup Resource Center, is a useful operator checkpoint. He confirmed a TCP/IP link from Bombay to UUNET over a 64-kilobit-per-second leased line and described a satellite network as a future expansion plan for places with poor data links. The email is evidence of what the project’s coordinator reported then; it is not a traffic measurement or proof that the proposed satellite system had already reached every institution. The 1998 ACM account later described ERNET as one of several distinct public or government connectivity systems, alongside NIC, STPI and VSNL.

Their users and purposes overlapped in places, but the organizations were not one national ISP under different names.

Public service changed the access boundary

India’s Department of Scientific and Industrial Research later summarized the distinction plainly: before the public launch, ERNET and the National Informatics Centre provided Internet access to people associated with educational institutions, research organizations or government departments. The Telecom Regulatory Authority of India dates the government’s launch of Internet services through Videsh Sanchar Nigam Limited (VSNL) to 15 August 1995. DSIR says the first public service covered six major cities.

Those official retrospectives support a careful formulation: India had Internet connectivity before the public subscription offer, but that connectivity was institutionally bounded.

The difference is not merely a wider customer list. A research network can build capability by concentrating on a group able to contribute expertise, equipment and local operations. Public service asks a different system to price access, connect to the telephone network, serve customers beyond the originating institutions and manage demand. ERNET did not simply become VSNL; the two played distinct roles. Nor did the public launch instantly remove access limits: the 1998 ACM account describes slow subscriber growth, metered local calls and a public network still shaped by telephone and gateway capacity.

Funding made continuity another question. The ACM authors reported that UNDP funding ended in 1995 and that ERNET then faced uncertainty about its future. That is a contemporary assessment from writers close to the Indian networking community, not proof that one grant decision alone determined the network’s fate. ERNET became an autonomous society in 1998, according to later institutional histories. The interval matters: an operational network, its project financing and its long-term institutional home were related but separate things to settle.

The evidence supports neither a tale of instant national access nor one of a solitary technical hero. ERNET’s importance lay in a group of institutions that built skills, mail exchange, gateways and domestic links together. Its first users could be connected to international research before a person in the general public could buy an Internet subscription. The achievement was real; so was the boundary around it.

Sources

Heng Lu’s Notes 64 and 65 inform the interpretation of voluntary institutional participation and running implementation; they are editorial lenses, not evidence about ERNET: https://heng.lu/minimum-initial-specification-localized-future-decision-voluntary-adoption-internet-coordination-system/ and https://heng.lu/running-code-primary-the-patch-needed-to-preserve-the-internet-original-design/.