Lu Heng · Why Internet power must change
A registry should keep the record. It should not control the network.
Every website, app and network depends on names and numbers that must remain unique. Today, central organisations keep those records and can turn a coordination role into power over the people who run the network. I argue for records that everyone can verify and an administrator that can always be replaced.
What is actually wrong?
A central record can become a central point of power.
When you type a domain name, a name system helps you reach the right website. When networks connect, IP addresses help them find one another. Someone has to keep these names and numbers unique and record who is using them. That basic coordination is necessary.
The danger starts when the organisation keeping the master record can decide who may receive, transfer or keep a number. A network operator may build the infrastructure, pay the bills and carry the risk, while the record keeper gains a veto over the network’s future. Bookkeeping has quietly become permission.
The same organisations may then claim to speak for a region, a continent or “the Internet community”. A service map and a meeting do not create political authority over everyone affected by the result. The people who operate networks are not automatically represented by the institution that records them.
This has already had a real-world warning. Before APNIC’s 2023 election, I raised serious concerns about its governance. Later disclosures showed that the organisation holding the Asia-Pacific number registry had, in practice, been under one person’s legal control for two decades. If a network has no guaranteed way to leave, an institutional crisis becomes an operational crisis. The cost reaches the network, its customers and its services.
Read the full argument
- Note 54Who Gets to Speak for a Continent, a Community, or “The End User” English original
- Note 72The Bill of Rights of Uniqueness Coordination English original
- Note 9On APNIC Governance and the Need for a Clean Break English original
What is my proposal?
Keep the record verifiable. Make the administrator replaceable.
Think about changing mobile providers: you can change the service without changing your phone number. Network operators need the same basic freedom. They should be able to move registration to a qualified successor without forcing customers to rebuild their systems around new addresses.
My proposal starts with a small set of technical rules that every participating network can verify. The shared record remains reliable, but no permanent gatekeeper gets to turn a notice into an order for the whole Internet. A qualified administrator can maintain the service, and another can take over when needed.
That is what I mean by decentralisation here: coordination remains possible, while the power to approve, refuse or rewrite the future is no longer held by one irreplaceable institution.
- Keep the common layer small
- Prevent conflicting numbers, protect the record and let compatible networks connect. Do not hide commercial or political preferences inside technical coordination.
- Let networks verify the record
- A network should be able to check why a record is valid without asking one permanent authority to approve the answer. Cryptographic proofs and distributed records can help.
- Make exit practical
- The record and essential services must be transferable. Operators can adopt later changes because they work, not because refusal brings institutional punishment.
Read the full argument
- Note 64Minimum Initial Specification, Localized Future Decision, and Voluntary Adoption for Internet Coordination Systems English original
- Note 13Decentralising Global IP Address Registration with Distributed Ledger Technology English original
- Note 72The Bill of Rights of Uniqueness Coordination English original
Why change the system instead of just the manager?
Replacing the people does not remove a permanent monopoly.
A better leadership team may improve an organisation for a while. It does not remove the underlying veto if the next team can still impose conditions, block a transfer or claim to represent people who never gave it that authority. The dependency itself has to change.
Moving the same power to a state or a larger global body would not solve the problem either. It would only create a larger gatekeeper. The durable answer is to separate the small technical job of keeping records unique from the political power to decide who may participate.
- Continuity cannot depend on goodwill
- Conflict, litigation or financial trouble can affect any institution. Critical networks need a tested alternative before the institution is in trouble.
- Technical coordination is not political representation
- A registry can keep a record accurate without speaking for a continent, a community or every end user.
- Disputes must not take down working networks
- Conflicting claims need an independent way to be resolved while unrelated customers and functioning services remain protected.
Read the full argument
Why must this happen now?
Build the way out before you need it.
An Internet address can sit inside customer systems, payment services, contracts and years of security settings. The longer a network depends on one gatekeeper, the more expensive a forced change becomes. A problem that looks like registry administration can reach ordinary businesses and daily life.
A real successor therefore has to be tested in advance: the records must remain accurate, another provider must be able to take over, and the security systems that verify address use must continue to work. Copying a database is not succession.
The pressure is increasing as address space becomes more valuable and Internet infrastructure becomes more politically important. My priority is simple: protect running networks and their customers, while making the organisation that maintains the record replaceable before a crisis removes the choice.
Read the full argument
Three places to continue
Start with representation, then see how a replaceable registry could work, and finally examine what must survive when an institution fails.
Who actually has a mandate to represent you?
Who Gets to Speak for a Continent, a Community, or “The End User”Note 54 · English original
How do rules work without a permanent authority?
Minimum Initial Specification, Localized Future Decision, and Voluntary Adoption for Internet Coordination SystemsNote 64 · English original
What must survive an institution’s failure?
The Registry Continuity Fallacy: Protect the Ledger, Not the GatekeeperNote 70 · English original
Explore the full reading path
How the arguments connect
Note 65 brings these seven Notes together: how registry power expands, who bears the cost, and how that arrangement could change.
Read the full argument
- Note 52On When Registry Power Detaches from Liability: Why the Present RIR Coordination Model Cannot Survive in Its Current Form English original
- Note 53On Internet Number Resources Are Not Political Property English original
- Note 56On Regional Internet Registries’ Thick Governance Turns Uniqueness into Double Extraction English original
- Note 58From Double Extraction to Sovereignty Inversion: How Nations Lose Sovereign Control to RIRs for US$100 English original
- Note 59The Poverty Penalty: How the RIR Model Taxes the Poor While Calling It Equality English original
- Note 61Running-Code Betrayal: How the RIR System Turned Consensus Against the Technical Community English original
- Note 62Mandate Laundering: From RIR Fantasy to Transition Architecture English original
Read the full argument
- Note 72The Bill of Rights of Uniqueness Coordination English original
- Note 65Running-Code Primacy: The Patch Needed to Preserve the Internet’s Original Design English original
- Note 45On the Manufactured Narrative of IPv4 Scarcity English original
- Note 1On AFRINIC, RIR Governance, and the Myth of “Community Ownership” English original
- Note 9On APNIC Governance and the Need for a Clean Break English original
- Note 37On Why NRS Exists — and Why Decentralization Is No Longer Optional English original
- Note 66On Why i.LEASE Exists — and Why the Broker Question Is Really a Registry-Risk Question English original
- Note 8Why Centralised Alternatives Fail, and Why a Decentralised Registry Is the Only Viable Path English original
- Note 13Decentralising Global IP Address Registration with Distributed Ledger Technology English original
- Note 54Who Gets to Speak for a Continent, a Community, or “The End User” English original
- Note 64Minimum Initial Specification, Localized Future Decision, and Voluntary Adoption for Internet Coordination Systems English original
- Note 70The Registry Continuity Fallacy: Protect the Ledger, Not the Gatekeeper English original
What I am working on.
These arguments also shape the businesses I build, the changes I advocate, and the public discussion I take part in.
Help shape what comes next.
I will keep writing about the problems I encounter and the changes I believe they require. If you want the Internet’s future to remain with the people who build and use it, follow my work and read the next Note.
Follow me on LinkedIn