Summary

  • The Soviet Union’s dissolution did not remove .su from the DNS root: the code remained delegated while Russia acquired .ru and the registry changed how new names could be registered.
  • Contemporary records distinguish root-zone status from second-level enrollment, surviving subdomains and actual network connectivity; none of those alone proves broad public access.

Analysis

A domain was already being assembled

At RIPE’s sixth meeting in August 1990, participants reported that the Soviet Union’s .su domain was being registered by the Unix User Group. Finnish and Dutch research networks, FUNET and CWI, were helping with secondary servers. The minutes said formal registration by the Network Information Center was expected in the coming days. IANA’s root-zone record gives 19 September 1990 as the registration date. Those records show both the practical work of building a namespace and the distinction between a meeting report and the date later recorded in the root registry.

The domain was not a national Internet connection. In a 1992 account, Relcom described network activity across the former Soviet republics while noting that most links were still UUCP and that its international connection ran by UUCP through Finland. Domain coordination, packet connectivity and public service were different layers. A name ending in .su could exist without proving that every place represented by the suffix had a working Internet path.

A state changed; the root record did not move on the same day

RIPN’s institutional history says administration of .su was entrusted to the newly established institute in 1993. Russia’s .ru zone followed in 1994, after providers agreed on its administration. The arrival of .ru did not itself replace every .su name. RIPN’s account says new second-level registrations in .su were stopped as the new Russian zone appeared, while third-level names continued to grow. It reports 30,000 hosts in .su by 1998. A 1998 RIPE list describes the older .su domain’s existing subdomains as managed by .ru.

These statements describe an administrative transition, not a single transfer of every name. The DNS root could continue to delegate .su while lower-level registration and maintenance changed. A zone could be expected to retire while existing names still resolved. The same label therefore carried both continuity and a policy question: should a namespace created for a state that no longer existed remain available to the communities and registrants already using it?

The expectation of retirement met an operating zone

RIPN’s history says an attempt to reopen general .su registration began in 2001, encountered a high proposed price and was then suspended at the request of Russia’s communications ministry. It dates priority registration for trademark holders to December 2002 and the start of open second-level registration to June 2003. These details come from the registry’s retrospective and should be read as its account of the sequence.

By June 2007, ICANN’s Governmental Advisory Committee minutes described .su as anticipated to retire and be replaced by .ru. In the same passage, the minutes recorded that .su was still active and accepting registrations. That is not evidence of a binding retirement order. It is evidence that an institutional expectation and an operating namespace could coexist in the record without being the same thing.

What persisted

The .su story is not proof that a country-code domain confers statehood, or that a root-zone record controls every aspect of a national network. It shows something narrower and more useful: political succession, code status, root delegation, registry administration, enrollment rules and reachability can change on different schedules. The IANA record gives the root’s present entry and original registration date; archival operational evidence shows names and network coordination continuing after the Soviet Union’s dissolution.

For a registrant, the distinction was concrete. A retained delegation could keep old addresses in the shared naming system; a registration freeze could limit new names without deleting existing ones; a later policy change could reopen enrollment. But a name’s continued presence in DNS did not establish that a host answered, that a service was maintained, or that its users were broadly connected. The historical record supports persistence of the namespace, not a universal story of access.

Heng Lu Note 64 informs only the editorial question of which rules must be shared and which later choices remain local or depend on adoption. It is not evidence about the decisions of SUUG, RIPN, Russian providers or ICANN participants.

Sources