Summary

  • The .br country-code domain was delegated on 18 April 1989, twenty-one months before FAPESP began exchanging Brazil's first international TCP/IP traffic in January 1991; the country already had academic links, but they used BITNET, HEPnet and related systems.
  • Demi Getschko's record connects that early act of anticipation to a longer operating duty: define a useful namespace, keep its authority reachable, distinguish delegation from ownership, and move responsibility from a trusted team into institutions that can outlast the original people.

The first fact in the history of .br is not a web page. It is a line of delegated authority. IANA's root record dates the country-code domain to 18 April 1989. The Brazilian network that most readers would recognise as the Internet came later: Demi Getschko's own historical account places the first international IP traffic on the FAPESP–Fermilab line in January 1991.

That gap of roughly twenty-one months is more revealing than a ceremonial “first.” Brazil was not disconnected in 1989. Researchers already used international academic networks. FAPESP had a 4,800-bit-per-second link to Fermilab carrying BITNET and HEPnet traffic; LNCC in Rio de Janeiro had made an earlier international connection. Email travelled, institutions coordinated and a national research-network effort was forming. What had not yet arrived at FAPESP was the protocol suite that made those networks part of the Internet's common packet fabric.

The country code therefore preceded the traffic it would later organise. NIC.br's anniversary history says Jon Postel delegated .br to the group running academic networks at FAPESP. Getschko recalls that Postel judged the Brazilian community mature enough to administer it. That is a retrospective account of a human judgment, not a machine-generated entitlement. A two-letter string derived from a country code did not operate itself. Someone had to answer mail, maintain name servers, decide who could be named beneath it and remain accountable when the arrangement changed.

Getschko was one member of that team, not a lone inventor. His accounts repeatedly name colleagues, universities, laboratories and the wider research community. Yet his individual role is unusually legible because it crosses the technical and institutional record. The Internet Hall of Fame says he has served as administrative contact for .br since 1989 and played a critical role in defining the Brazilian DNS tree and registry rules. IANA's current delegation record still lists him as administrative contact, with CGI.br as the manager.

Continuity should not be mistaken for stasis. The job in 1989 was small enough for registrations to be made manually. The official anniversary account says .br initially identified machines in the academic environment. Getschko's 2006 history is more specific: on 18 April, the root already pointed to name servers at FAPESP. The namespace was an operational promise attached to actual machines and people, even though the surrounding Brazilian connectivity was still transitioning.

In January 1991, the same FAPESP–Fermilab line began carrying TCP/IP. The old academic systems did not vanish at midnight; the interview record says BITNET survived until 1997. Protocol migration overlapped with naming continuity. This matters because infrastructure rarely replaces one world with another in a clean cut. A durable identifier can bridge an untidy transition, allowing organisations to change transport without abandoning the names by which others find them.

The namespace itself then acquired structure. In 1991, universities and research institutes could sit directly under .br, while functional branches such as com.br, net.br, org.br, gov.br and mil.br were created. That was not merely clerical sorting. The shape of a DNS tree decides who applies where, which policies govern a name and which institution can change it. Taxonomy becomes a control surface when the names are globally relied upon.

Commercial use expanded at the end of 1994. In 1995 the Brazilian Internet Steering Committee, CGI.br, was created and assumed responsibility for the .br root and IPv4 distribution. Getschko's history notes an important form of succession: the functions were redelegated to the same operating team, now supported through the new governance arrangement. The authority moved into a broader institution without pretending that operational knowledge could be recreated instantly.

NIC.br later became the execution arm. CGI.br's official history says NIC.br has implemented the committee's decisions and projects since 2005 and operates the domain today. The sequence—FAPESP team, CGI.br policy authority, NIC.br operations—shows institutional layering rather than a single founding act. Each layer answered a different question: who can run the service, who can set legitimate rules, and what organisation can preserve the work at national scale?

Jon Postel's RFC 1591 supplies a useful vocabulary, with one chronological warning. It was published in March 1994, nearly five years after the .br delegation, so it cannot be presented as the written rule that governed the 1989 decision. It does, however, articulate the model that the Brazilian record came to embody. A designated manager supervises the namespace; the authority is a trustee for both the nation and the global Internet community; the emphasis belongs on responsibility and service, not ownership.

That distinction limits heroic history. If .br were treated as Getschko's possession, the long tenure of one administrative contact would look like concentration. Treated as a stewardship record, the test is different: did the service remain stable, did policy acquire legitimate institutional oversight, and could responsibility survive changes in technology and organisation? The present IANA record provides evidence of continuity, but it is not by itself a full audit of resilience, representation or succession.

Nor does the early delegation prove that every machine in Brazil could immediately use DNS over an IP path. It proves that the root recognised an authority and that designated name servers existed. The 1991 IP milestone proves a different thing: the shared transport became available. The later tree design, automated registration and governance changes prove still other things. Combining them into one “birth of the Internet” date erases the engineering work between designation, reachability, service and scale.

Getschko's contribution is most useful when understood as preparation under uncertainty. The FAPESP team obtained a national namespace before it could know the size of the Brazilian Internet, the commercial demand for names or the institutions that would eventually govern it. They made an irreversible commitment—the country code would become a stable point of reference—while leaving the internal tree and operating arrangements capable of change.

That is the modern lesson hidden in the chronology. Critical infrastructure is often named after it is built, as if identity were packaging. .br shows the reverse. A well-stewarded name can coordinate investment before the network is mature. It gives operators a destination for configuration, institutions a place in a shared hierarchy and users an expectation of continuity. But the advantage exists only if delegated authority is tied to reachable servers, documented policy, credible contacts and a succession path.

The country code arrived before Brazil spoke TCP/IP because a community was preparing to join a network whose eventual scale it could not yet see. The notable achievement was not guessing the future perfectly. It was creating an identifier whose technical and institutional custodians could keep adapting after the guess became real.

Sources