Decentralized Internet: Advantages, Challenges and Impact
What changes when a network can leave its registry? Explore the benefits, hard design questions and urgency of decentralized IP registration.
A useful test for a decentralized Internet is this: if the organization that registers your network’s addresses stopped working tomorrow, could you prove what you hold and keep operating?
Lu Heng’s case for decentralization starts with that dependency. Networks can be independently built and operated while still relying on a small number of institutions to recognize their essential resources. The question is how to remove that single point of control without losing the shared records that make a global Internet possible.
The benefit: an operator can leave without losing its network
Consider a provider serving homes and businesses. Its customers need continuity. They do not benefit from the provider being locked into one administrator for the lifetime of its address resources.
Under the approach Lu Heng advocates, the provider would be able to carry independently verifiable records to another registration service. Changing an administrator would not mean surrendering the resources around which it built its network.
That changes the relationship. A registry has to provide a useful service because the network has a meaningful choice. The right to leave makes accountability practical: the operator can act instead of waiting indefinitely for internal reform.
How a distributed registry would work
The proposal in Lu Heng’s Note 13 uses a distributed ledger for IP address registration. Several independent participants keep and verify a shared history of records. Cryptographic signatures help establish who can authorize a change. Common rules prevent two conflicting claims from both being accepted as the same resource holding.
The purpose is to make a resource record survive the failure or arbitrary decision of any one administrator. It is not enough to put copies of a database on several servers if one organization still has the sole authority to change it.
Registration must also work with real networks. A correct record does not route a packet by itself. Operators have to recognize the records and use compatible systems. The proposed registry therefore needs a transition that existing networks can join without breaking their connections.
The hard questions are part of the design
What happens if a holder loses its signing key? Who resolves two competing claims to the same resource? How does a valid court order take effect without giving one participant power to rewrite everyone’s records? Can participants replace a failed service provider and still agree on the history?
Those are engineering and governance requirements, not reasons to leave the current dependency untouched. A credible implementation must show its answers in operation. A ledger with a hidden master key would reproduce the problem. A system that cannot recover from an honest mistake would create a new one.
Decentralizing registration also does not automatically encrypt traffic, stop surveillance or make every service faster. Those require their own technical measures. The gain here is specific and important: reducing the power of a single institution over the recognition of network resources.
Why more participation is not enough
Inviting more people to a meeting may improve a decision. It does not, on its own, change who has the final ability to approve, revoke or withhold recognition.
Nor does moving that authority to a larger international organization remove the dependency. Note 8 sets out why Lu Heng rejects centralized replacements. The design must let independent networks coordinate without making one institution indispensable.
Why the change matters now
Every network built around resources it cannot take elsewhere adds to the cost of a future crisis. Preparing an alternative while systems are working leaves room to test, compare and transition. Waiting for an institution to fail takes that room away.
The ambition is a global Internet that keeps working through institutional failure, with resource holders able to choose the services they use. Continue with Lu Heng’s Note 37: why decentralization is no longer optional.