Summary

  • RFC 1401 prints the IAB's 30 March 1992 DNS appeal, DISA's 16 April reply and the IAB's 19 May follow-up. The IAB asked for views by 20 April and urged policymakers to set deadlines for moving connected systems to DNS.
  • DISA answered on time with a promise to study technical and economic effects on Defense Department systems and report later decisions. The printed correspondence records neither an adoption decision nor a completed transition.

Colonel Kenneth Thomas replied four days before the IAB's requested date. That fact sounds like progress until one reads what he promised: DISA would staff the DNS question, examine its technical and economic impact, and tell the IAB the resulting decisions later. The three letters printed in RFC 1401 end with Lyman Chapin thanking him for the study and discussing a possible meeting. The calendar obligation was met. The operational question stayed open.

The IAB's problem was concrete. Its attached February 1992 memorandum argued that a single centrally maintained host-name file could no longer keep pace with the Internet's size and changing records. It pointed to remaining MILNET reliance on hosts.txt, despite a transition plan worked out in 1987. Different communities could then answer the same name with inconsistent data. The IAB warned that a user would experience this as a failure to reach a resource, not as a debate about database architecture.

Mail made the loss sharper. DNS MX records could name and prioritize alternative relays when the final destination was unavailable. A plain host table did not carry that behavior. The IAB proposed a grudging fallback for communities still tied to tables: bring DNS information into the NIC dataset, keep it synchronized, and at minimum include MX data. That was a compatibility repair, while its preferred policy was uniform DNS use with deadlines set by authorities that could impose them.

The older RFC 1031 explains why the 1992 letter had teeth. In 1987 it had described a potential MILNET transition timetable ending in October 1989, while opening with the admission that no official policy had yet been published. Table-only, mixed and DNS-only hosts could coexist during migration. Five years later, the IAB was still writing about sizeable pockets of table dependence. Its account is an institutional warning in a letter, not a measured census of every MILNET host.

The exchange also locates the lever. The IAB could publish a technical case and urge deadlines; DISA had to weigh what adoption would mean for DoD systems. Thomas's answer recognized performance and interoperability but committed to a review. The third letter adds no implementation order, date or result. An RFC can make the request and the reply public. It cannot turn a study promise into a configured resolver.

There is a precise uncertainty here: the three printed letters contain no later DISA decision or evidence that the proposed meeting took place. The record still shows a useful institutional fact. One body could describe the operational cost of incompatible names; another held the next decision. Between them, a timely letter was easy to produce and a common naming system was not.