Summary

  • In June 2014, holders of seven terrorism judgments served ICANN with writs and subpoenas aimed at .ir, .sy, .kp, corresponding internationalised domains and supporting IP-address information.
  • The district court found the ccTLDs non-attachable under District of Columbia law. It expressly did not decide that a ccTLD could never be property.
  • The D.C. Circuit affirmed on narrower and alternative grounds. The named Weinstein claim was forfeited before the merits; for four preserved judgments, the court assumed property and sovereign ownership without deciding either, then protected non-liable third-party interests under 28 U.S.C. §1610(g)(3).

A sale without the operational asset

The proposed remedy sounded familiar. Attach an asset, sell it and apply the proceeds to a judgment. The object was not a building or an account. The creditors wanted .ir management delegated to them so they could sell or license its operation to a third party.

That formulation concealed what the purchaser would need. A root-zone entry could point resolvers toward new authoritative servers. It would not contain the second-level registrations held in the existing .ir registry. Without that dataset, names already used by businesses, institutions and individuals could stop resolving through the changed root. The outgoing manager was abroad. The D.C. Circuit saw no reason to suppose a United States court could compel it to hand over the database.

The writ had reached a United States corporation. The usable namespace had not thereby entered the courtroom.

Seven judgments, one set of writs

The creditors were victims of terrorist attacks and surviving relatives. Between 2002 and 2012, the seven groups had obtained judgments worth hundreds of millions of dollars against Iran, Syria or North Korea. Their efforts to collect had repeatedly encountered the separate immunity rules governing sovereign property.

On 23 and 24 June 2014, they served ICANN with seven writs of attachment and subpoenas. The requested targets included .ir, .sy and .kp, their internationalised counterparts and “supporting IP addresses.” The first subpoena’s Schedule A asked for documents about contracts, allocations, licences, assignments, transfers, payments and communications involving Iran and its security ministry.

Those verbs were the creditors’ discovery formulation. They did not prove that ICANN had licensed a sovereign-owned asset or possessed a transferable title. ICANN answered that it was not indebted to the defendants, held none of their goods, chattels or credits, and had no agreements with or contributions from the three ccTLD managers. It moved to quash.

What ICANN could do in 2014

The appellate opinion separated the then-current root-zone chain into three acts. ICANN screened a delegation or redelegation request and made a recommendation. The United States Department of Commerce authorised the change. Verisign implemented it in the root-zone file. The United States called its role “largely symbolic,” limited to checking process and technical errors. That was the government’s characterisation; the formal approval step still appeared in the chain described by the court.

The chain was consequential but incomplete. The ccTLD manager operated authoritative servers and maintained the registry below the top level. Internet service providers and users followed the established root because it gave consistent answers and benefited from network effects. A state could also direct networks within its borders toward a different root. Formal approval, technical publication, registry custody and network adoption were separate control surfaces.

RFC 1591 supplied a service vocabulary for that arrangement. It described a designated manager as a trustee for the nation and global Internet community, demanded equitable and technically competent operation, and said concerns about rights and ownership were inappropriate. It was, and is, an informational RFC—not an Internet Standard, a property statute or proof that every affected person authorised the institution speaking in the community’s name.

The district court kept its holding narrow

On 10 November 2014, the district court quashed the writs through Federal Rule of Civil Procedure 69 and District of Columbia Code §16-544. That local provision permitted attachment of a debtor’s goods, chattels and credits. Drawing on a Virginia domain-name decision, Judge Royce C. Lamberth found that the ccTLDs could not be conceptualised apart from the continuing services that made them operational. A creditor could not be inserted by garnishment into that arrangement.

The footnote is as important as the result. The court said its conclusion did not mean ccTLDs could never be property. It meant they were not attachable property within this statutory scheme. Because that disposed of the writs, the court denied the creditors’ request for extended discovery as moot.

The D.C. Circuit later expressed reservations about using local attachment law for the federal question. It nevertheless assumed, for the analysis that followed, that local law did not stop the creditors. The affirmance therefore did not turn the district court’s service reasoning into a universal rule of ownership.

The caption obscures the procedural split

The consolidated appeal carries the name Weinstein, but the Weinstein judgment did not reach the final third-party-interest analysis. Weinstein, Haim I and Stern had been entered under the former terrorism exception, 28 U.S.C. §1605(a)(7), and had not been converted to the current §1605A route. The appellate court held that their collection claims were forfeited in toto.

Haim II, Rubin, Wyatt and Calderon-Cardona preserved only the terrorism exception in §1610(g). The commercial-activity theory was forfeited. TRIA had not been raised in the district court. Arguments about IP addresses were waived on appeal. The opinion therefore established no ownership or attachment rule for IP number resources.

This sorting determines the strength of every later sentence. Only four judgment groups reached §1610(g)(3). For them, the court assumed without deciding both that the ccTLDs were property and that the defendant sovereigns held some attachable ownership interest. An assumption used to test the next question is not a holding.

The people outside the judgments

Section 1610(g)(3) preserved a court’s authority to prevent impairment of an interest held by a person not liable in the action that produced the judgment. The D.C. Circuit found enormous third-party interests and no way to execute without impairing them.

A prospective .ir manager needed technical competence and a commitment to the relevant user population. Even a qualified successor would begin with the root pointer, not the live registration ledger. The current manager’s cooperation was therefore central. Existing registrants had not incurred Iran’s judgment debt, yet their reachability would be placed inside the enforcement experiment. Networks outside the United States were also not compelled to accept the replacement.

The court considered a possible split root. If foreign ISPs continued querying the old .ir servers while other networks followed the changed root, identical names could yield different destinations. ICANN and the United States described severe consequences for stability and the prevailing governance model. The court did not decide whether their “doomsday” description was accurate. It said ICANN’s own protected third-party interest was sufficient, then reasoned that a forced court-directed change could give networks an incentive to turn away.

That distinction prevents institutional rhetoric from becoming a finding. The court protected a legally recognised interest. It did not constitutionalise ICANN, declare the current root metaphysically unique or hold that every challenge to the coordinator would break the Internet.

Who paid, who gained and what was locked

The creditors carried years of collection delay and litigation expense; the record here does not quantify it. ICANN bore compliance and defence costs. The ccTLD managers faced displacement and data-transfer demands. Registrants, users and networks bore the potential cost of broken resolution, inconsistent answers or migration. None of those third parties was liable on the terrorism judgments.

The immediate beneficiaries of the ruling included the existing managers and users whose operation remained unchanged, and ICANN, whose place in the root process was preserved. The debtor states also escaped this execution route. That consequence should not be confused with the court’s stated justification: procedural limits and protection of non-liable interests, not a finding that the judgments lacked merit.

The lock-in was real but not absolute. Contracts and the published root concentrated execution. Registry data remained elsewhere. Network effects made compatibility valuable. An alternative root remained technically possible but economically and politically costly. ICANN’s effective power came from the alignment of these layers, not from a judgment declaring ownership.

A narrower enforcement path

A credible counterfactual would first search for a separable debtor-specific asset: a verified receivable, payment or contractual right actually held by a reachable garnishee. If such an asset existed, a court could address it without auctioning operation of the live namespace. This record does not establish that it existed.

Any attempted transfer of management would require more than a writ to ICANN. It would need authenticated registry-data escrow, an independent competence review, protection of existing registrations, a record of affected-party support and objections, staged testing and rollback. Those conditions would make the costs visible before an irreversible cutover. They do not prove that the court had authority to order the plan.

For number-resource holders, the case is an architectural warning rather than a property precedent. A registry entry, a legal title, an RPKI or DNS assertion, an operational database and global routing or resolution acceptance are different things. Before enforcing or defending a resource claim, map who can adjudicate, who can sign, who can publish, who holds the ledger and who can keep the service running. Community, consensus, stewardship and stability are questions to test, not evidence of mandate.

Sources