Why IP addresses are economic infrastructure

Imagine a company moving its order-processing system to a better network provider. The new service works, but a supplier stops accepting its connections. Nothing is wrong with the orders: the supplier’s security rules still recognise the old IP address.

Fixing that rule may be simple. Finding every customer, partner and system with the same dependency may not be. The company has changed a technical setting, but the work reaches beyond its own network.

This is why IP addresses belong in a discussion of economic infrastructure. The important question is what people have built around them, and whether that investment can survive a change of provider or administrator.

What an address does, and what grows around it

An IP address helps networks deliver data to a destination. Some addresses are temporary and easy to replace. Others become embedded in the way organisations communicate.

A supplier may allow connections only from a known address. An application may use a fixed destination. An operations team may have years of monitoring and incident records attached to an address range. None of these makes an IP address proof of a person’s identity. They do make changing it a coordination problem.

Think of the distinction as the number itself and the relationships built around the number. Acquiring another number does not automatically move those relationships.

The cost of changing the number

In Note 68, Lu Heng develops the economics of network identity. His focus is the accumulated dependence on an address: who already recognises it, what would need to change, and who would have to do that work.

This gives a practical way to distinguish ordinary address capacity from continuity. A short-lived test server may be easy to renumber. A service connected to many independent partners may be much harder. The difference lies in the consequences of moving, not simply in the price of an address block.

That difference also changes competition. A customer who can keep its working relationships has more freedom to choose a better network. A provider can earn that customer’s business through service quality, rather than through the difficulty of leaving.

Registration is not a mandate to govern the economy

Networks need dependable records so that the same globally unique identifiers are not assigned incompatibly. They also need to know who can demonstrate control of a resource. These are important shared functions.

RFC 7020 distinguishes number registration from routing: accurate registration is a registry responsibility; whether and how networks announce addresses is an operational matter. A registry is not the central switch through which every connection passes.

Yet control of a record can still have economic force. If working services depend on that record, an administrative decision can affect customers who have no direct relationship with the administrator.

Lu Heng examines this change in Note 42, on the power acquired by a once modest bookkeeping role. His objection is that growing dependence has not produced equally strong accountability or a practical ability to replace the institution.

Calling addresses “infrastructure” should therefore sharpen the limits on administrative power. The importance of a service is a reason to make it dependable and replaceable. It does not, by itself, give its administrator political authority over the people and businesses using it.

More addresses do not settle the question of power

IPv4 and IPv6 provide different address spaces. IPv6 uses 128-bit addresses, allowing a much larger space than IPv4. That technical difference matters when designing networks.

It does not answer who controls an address record, whether an operator can replace an administrator, or how a business preserves its established connections. A larger supply of identifiers and a more accountable coordination system solve different problems.

The same distinction applies to buying or leasing IPv4 capacity. Either may meet an operational need. Neither transaction, on its own, demonstrates that the network is free of registration dependencies or can move its identity to every provider. Start with the service that must keep working, then examine the dependencies behind the proposed arrangement.

Keep coordination; give operators a real way to leave

In Note 72, the Bill of Rights of Uniqueness Coordination, Lu Heng proposes limiting the common system to essential functions: unique identifiers, accurate records, evidence of control, security assertions and traceable changes.

Portability is central to the proposal. An operator should be able to replace a failing administrator without sacrificing the resources on which its network depends. That requires a credible transition: other participants must be able to verify the records and continue using the network throughout the change.

This is a proposed direction for Internet coordination, not a claim that every registry already offers such an exit. Its test is concrete: can an alternative preserve the shared function while removing the permanent dependence on one organisation?

Why preparation matters before the next move

Consider the order-processing service again. Its next migration becomes easier if the team knows which external rules depend on its addresses, which records support its routes, and what a replacement provider would need to carry them.

Waiting until a contract ends or a service fails makes those discoveries harder. Every new integration can add another dependency. Establishing a workable alternative takes cooperation from people outside the organisation; it cannot always be improvised during an outage.

A useful first step is to map one service:

1. Which addresses do its users and partners depend on?
2. Who supplies the network connection, and who maintains the relevant records?
3. What can change independently, and what would force other organisations to change their systems?
4. Has a replacement path been demonstrated while the service is still running?

Lu Heng presents LARUS One as a practical application of separating lasting network identity from its delivery provider. The wider institutional question remains: the coordination beneath a service must also have a credible replacement path.

Continue with Note 68: the economics of network identity for the original argument, or the guide to mapping governance risk to examine another service. The aim is an Internet where continuity gives people more choice, and essential coordination cannot become a permanent claim to rule.