SRI-NIC: The First Name Authority

How the Stanford Research Institute's Network Information Center became the internet's first naming authority.

Every system needs someone in charge. For ARPANET naming, that someone was SRI-NIC — the Network Information Center at Stanford Research Institute. Understanding SRI-NIC’s role helps explain both why HOSTS.TXT worked as long as it did, and why DNS would eventually need to distribute that authority.

The Birth of NIC

In 1969, when ARPANET was first coming online, the Defense Advanced Research Projects Agency (DARPA) needed someone to manage the network’s administrative functions. Stanford Research Institute (later SRI International) won the contract.

Their responsibilities included:

  • Maintaining HOSTS.TXT — the master file of all host names and addresses
  • Assigning network numbers — coordinating address allocation
  • Publishing documentation — RFCs, protocol specifications, network maps
  • User support — helping new sites get connected

The NIC was, in essence, the internet’s first bureaucracy. And like any bureaucracy, it had both benefits and limitations.

How SRI-NIC Operated

The Update Process

Adding or changing a host name followed a specific procedure:

  1. Email Request — Site administrators would send an email to HOSTMASTER@SRI-NIC requesting a change
  2. Manual Review — NIC staff would review the request for conflicts and format compliance
  3. File Update — Approved changes were manually added to HOSTS.TXT
  4. Distribution — The updated file was made available for FTP download

This process typically took 1-3 days, though it could be faster for urgent requests or slower during busy periods.

The People Behind NIC

Several key individuals kept the system running:

  • Elizabeth “Jake” Feinler — Director of NIC from 1972 to 1989, she oversaw the transition from HOSTS.TXT to DNS
  • Ken Harrenstien — Wrote many of the host table specifications
  • Vic White — Key contributor to NIC operations

These people handled an ever-increasing workload with limited resources, maintaining reliability as the network grew by orders of magnitude.

The WHOIS Service

SRI-NIC also created the first WHOIS service — a protocol for looking up information about network resources and contacts. This precursor to modern WHOIS databases let users query:

$ whois SRI-NIC
Host: SRI-NIC
Address: 333 Ravenswood Ave, Menlo Park, CA 94025
System: DEC-2060
Contact: HOSTMASTER@SRI-NIC

RFC 812 (1982) formalized this service, which would later evolve into the WHOIS protocol we still use today for domain registration information.

Scaling Challenges at NIC

As the network grew, NIC faced mounting challenges:

Staffing

The same small team that handled dozens of hosts was now handling hundreds. Email volume to HOSTMASTER increased exponentially. Processing requests became a full-time job for multiple people.

Infrastructure

The SRI-NIC server itself became a bottleneck. Every host on the network periodically downloaded HOSTS.TXT from this single machine. By the early 1980s, this FTP traffic was consuming significant bandwidth.

Policy Disputes

As the network became more valuable, naming disputes became more contentious:

  • Who should get desirable short names?
  • How should new organizations be named?
  • What about international sites?
  • Who settles disagreements?

NIC had to make judgment calls that not everyone agreed with, but there was no formal appeals process.

Distributed Authority

Perhaps the biggest limitation: NIC couldn’t delegate. Every name change, no matter how minor, required NIC involvement. There was no way to say “MIT manages all MIT-* names” — NIC had to process every request.

The Numbers Game

SRI-NIC’s bandwidth consumption tells the story:

Year HOSTS.TXT Downloads/Day File Size
1980 ~200 ~20 KB
1982 ~500 ~50 KB
1984 ~1,000+ ~100 KB+

Each download transferred the entire file. Even with incremental improvements, this linear scaling couldn’t continue.

Seeds of Change

By 1983, several factors were converging:

  1. DARPA Involvement — The defense network’s sponsor was pushing for a more scalable solution
  2. Research Proposals — Multiple groups were exploring hierarchical naming systems
  3. International Growth — Non-US networks were joining, requiring new naming conventions
  4. TCP/IP Adoption — The January 1983 “flag day” transition to TCP/IP would connect even more networks

Jon Postel at USC’s Information Sciences Institute coordinated much of this effort. Working with DARPA and SRI-NIC, he helped shepherd the proposals that would become DNS.

The Handoff

When DNS launched in 1984, SRI-NIC didn’t disappear. Instead, its role evolved:

  • NIC became the root zone administrator for DNS
  • NIC managed the .arpa domain (for legacy compatibility)
  • NIC continued WHOIS services (now for domains instead of just hosts)

The transition was gradual. HOSTS.TXT was maintained alongside DNS for years, eventually being deprecated in 1989.

Key Takeaways

  • SRI-NIC was the internet’s first naming authority, operating from 1969 to 1989
  • NIC maintained HOSTS.TXT, assigned network numbers, and provided user support
  • The WHOIS service originated at NIC and evolved into modern domain registration lookup
  • Scaling challenges — staffing, infrastructure, policy, and inability to delegate — made centralized authority unsustainable
  • NIC’s role evolved with DNS, transitioning to root zone administration

Legacy

SRI-NIC established many patterns that DNS would inherit:

  • Centralized policy with distributed implementation
  • Contact information registration (WHOIS)
  • Formal request processes for changes
  • Documentation and RFC publication

Understanding NIC helps us appreciate both the genius and the limitations of DNS design. The new system had to preserve what worked while fundamentally changing what didn’t.

Next

With the historical context established, we’re ready to explore how Paul Mockapetris and Jon Postel designed the Domain Name System. The next section covers the invention of DNS and the key decisions that made it work.