A satellite quantum network needs a memory budget as well as an orbit
QNETSIM asks whether moving links, stored states and classical messages line up to deliver a usable service.
A satellite passing within reach of two ground stations does not by itself establish a usable quantum-network service. QNETSIM, described in an October 27 preprint, models orbital motion alongside optical links, quantum memories and the classical coordination needed to complete network operations.[1]
The framework distinguishes geometric contact from delivered service. It tracks storage ages and resource availability, separates raw delivery from fidelity-qualified delivery, and distinguishes a physically ready resource from one whose required classical information is available. Its results describe simulated network designs, not a deployed constellation.[1]
Consider a deliberately simplified example, not a QNETSIM result. One link supplies a resource at 20 milliseconds, but the second arrives at 140 milliseconds. If the first resource remains usable for only 100 milliseconds after arrival, the overlap has been lost. Counting two successful links would hide that failure to complete the combined operation.
A queue introduces another distinction. A request might be served quickly because the system already holds a suitable resource; starting with empty memories could take much longer. The paper therefore separates stocked-service latency from a cold start.[1]
Those distinctions give a network planner more informative questions than how many links exist on a map. When does a request enter the system? What is waiting in memory? Which stored resource is consumed next? Can the receiving application use the result by its deadline? A reported delivery rate needs that operational context before it becomes useful for a service decision.
A public repository contains QNETSIM code and application launchers, including ground and satellite configurations. The repository was inspected for this report but not executed, and a software license was not verified. Public visibility should therefore not be described here as a verified open-source release.[2]
The most useful next comparison would keep the application request and its quality threshold fixed, then change the orbit, memory lifetime or scheduling policy one at a time. That would show which investment actually improves completed service, rather than merely producing a busier simulation.