Exploring the Landscape of Internet Governance
Who can decide for the networks that keep us connected? Understand Internet governance, its authority problem and Lu Heng’s proposed direction.
A room full of people can discuss how the Internet should work. That does not tell us who gave them permission to decide for everyone else.
Start with a local Internet provider. It buys equipment, connects homes and answers the phone when service fails. Now imagine a decision made elsewhere changes how other networks recognize its IP addresses. The provider and its customers carry the consequences. Who authorized the decision, and who is responsible if it goes wrong?
That is the question Lu Heng brings to Internet governance. Understanding his argument starts with separating three things: doing useful technical work, taking part in a discussion and having authority to decide for another party.
What is Internet governance?
Internet governance covers the decisions and arrangements that shape how the Internet works: technical standards, names and numbers, network operations, public policy and the rules of online services. There is no single organization running all of it.
The roles differ. Network operators connect their systems and carry traffic. The IETF develops voluntary technical standards; it does not run or police the Internet. IANA functions coordinate domain names, number resources and protocol parameters, with different arrangements for each. Five Regional Internet Registries—AFRINIC, APNIC, ARIN, LACNIC and RIPE NCC—provide number-resource services in their regions.
A forum has a different job again. The Internet Governance Forum brings people together for discussion; it does not negotiate binding decisions. Governments and courts exercise public authority within their jurisdictions. Calling all these activities “governance” does not make their powers interchangeable.
Joining a discussion is not the same as authorizing a decision
The multistakeholder approach invites governments, businesses, engineers, civil society and others into the discussion. Their knowledge and objections can improve a decision. But someone affected by a decision is not automatically authorized to make it on another person’s behalf.
In Note 73, Lu Heng argues that the model helped avoid some forms of state control, but then allowed participation to stand in for a mandate. His objection is not to hearing more voices. It is to treating a room of participants as if it were everyone whose interests they discuss.
A representative might have authority from an employer, a membership or a government. The scope of that authority matters. Serving networks across a continent does not, by itself, authorize an institution to speak for the continent. An advisory committee does not become the voice of every Internet user simply by adopting that description.
Who bears the cost when the decision fails?
Return to the provider connecting those homes. A delayed transfer or disputed resource record can mean engineering work, postponed investment and uncertainty for customers. The people making the administrative decision may not face the same losses.
Lu Heng calls this an agency problem in Note 32: an administrator can gain influence by expanding its role while operators bear the cost of that expansion. A well-attended meeting does not correct this mismatch.
This does not mean a registry controls every router. It means that control over records on which networks rely can have economic and operational consequences. Those consequences are the reason to ask how far the administrator’s powers should extend.
The solution: make coordination useful and replaceable
Some work must be shared. Address records need to be consistent. Conflicting assignments must be detected. Other networks need a dependable way to verify who controls a resource.
Lu Heng’s proposed direction keeps that common work narrow. Operators should have verifiable records, portable recognition of their resources and a way to replace a failed registration service while keeping their networks running. The administrator serves the system; the system should not depend on preserving one administrator indefinitely.
The distinction is practical. Inviting more people to a meeting changes who can speak. Portability changes what a resource holder can do when the service stops working. Both may matter, but one cannot substitute for the other.
A new global authority would retain the same dependence at a larger scale. The alternative developed in these Notes is decentralized coordination with common standards, rather than a different institution holding the same unchecked power.
Why make the change before a crisis?
Customers still need their connections when an institution is in trouble. Operators cannot suspend every service while a governance dispute runs its course. Replacement and continuity arrangements have to be tested while the existing system is still working.
The case for change therefore rests on everyday reliability as well as principle. Can the records survive institutional failure? Can a provider move without abandoning the network built around them? Can an administrator be held responsible for the power it exercises?
A better question to take into the next discussion
Ask who can authorize this particular decision, who carries its cost and what happens if the administrator fails. That is more revealing than counting how many kinds of stakeholder appeared in the room.
For the argument behind those questions, read Lu Heng’s Note 73, on how participation became a claim to govern.