Internet governance in 2026: why decentralisation matters
Who can decide how a network operates, and what happens if that authority fails? An introduction to Lu Heng’s case for replaceable coordination.
A network can buy its equipment, build its connections and serve its customers, yet still depend on an outside administrator to keep its address records usable. What should happen if that administrator fails—or starts making decisions far beyond keeping the records accurate?
This is where the case for a decentralised Internet becomes concrete. The question is not simply how many people attend a policy meeting. It is whether the networks doing the work have a dependable way to keep operating.
What needs to be shared?
Think of an address book used by many separate delivery companies. They need consistent addresses to find the right destinations. They do not need the person maintaining the book to run every company.
On the Internet, IP addresses identify network interfaces; routing systems help networks forward traffic toward its destination. Public address assignments need coordination so that unrelated holders are not given conflicting claims to the same resource. That is different from requiring every device, including devices behind a home router, to have its own public address.
IANA coordinates the global pool of Internet number resources, working with five Regional Internet Registries. The Internet Numbers Registry System describes how these responsibilities are distributed. Registries maintain records; network operators decide how to connect and carry traffic. Those functions interact, but they are not interchangeable.
When does coordination become control?
A record can become a point of leverage when other parties rely on it and its holder has no workable alternative. The administrator need not own the cables or operate the routers to make a transfer, expansion or recovery difficult.
Lu Heng challenges the jump from providing that service to claiming a broader right to govern. Serving a region does not by itself authorize an institution to speak for everyone living there. Inviting businesses, governments and civil society into a discussion does not resolve who authorized a binding decision.
In Note 73, on the multistakeholder model, he separates participation from mandate. Listening to more people can improve technical judgment. It cannot automatically turn a meeting into a representative government.
What would decentralisation change?
The test is practical: could a network replace a failed registration service while preserving reliable recognition of its resources?
Lu Heng's proposed answer keeps a small set of shared functions: unique assignments, credible proof of control, accurate records, auditable changes and reliable security information. It adds a route to replacement. Decisions that do not need to be common remain closer to the people operating the networks.
This is a design direction, not a claim that universal registry portability is already available. Changing a database entry alone would not make every network accept the change. A replacement has to preserve consistent records and the operational relationships that rely on them.
Why a larger committee does not solve the same problem
More consultation can reveal mistakes. Better transparency can expose a bad decision. Neither gives an operator a way out if the essential service becomes unavailable.
Nor would replacing today's institutions with one stronger global authority remove the dependency. The important change is the ability to replace an administrator while preserving the useful function it performs.
Why prepare before something breaks?
Customers still need their connections while a dispute or institutional failure is being resolved. An operator cannot invent and test a continuity arrangement after everyone has already lost access to the service it needs.
That gives the argument its urgency: build and test the replacement path while the existing system still works. The aim is a network whose continuity does not depend on the permanent health of one organization.
Continue with how this dependency becomes a business risk. For Lu Heng's proposed principles in his own words, read the Bill of Rights of Uniqueness Coordination.