Here is a theory around the reasons for the #Starlink outage we saw a few hours ago. There is an entity in the Starlink system I know as “The Central Scheduler”. Sounds grandiose, I know, but it’s the “brain” of the system.
The scheduler is tasked with orchestrating all the resources in the space segment, such as gateway connections, spot beam allocations, ISL interconnections, and likely a few others. It re-calculates the constellation in 15-second blocks.
We also know SpaceX was planning to implement a FUP starting in December. The purpose of a FUP is to control an ISP’s top consumers’ consumption of the most expensive parts of the network path, so that 10% of costumers are eating 90% of resources.
What is Starlink’s most expensive resource? I’d argue the space segment. You cannot easily just light up a few more fiber cores to meet increasing demand. Even by launching more satellites, Starlink won’t be able to meet heavy demand in congested cells - Gen2 should tackle that.
As of today, Starlink has in orbit ~74% of the complete Gen1 constellation, some of which may not be operational yet. It is also still ramping up gateway capacity to support the increasing traffic loads, ISL, and the maritime / aviation markets.
The only effective way to apply a FUP that conserves space segment resources is to limit the bandwidth assigned to a specific terminal, via OFDMA subcarriers and time slots. The element that controls this is the satellite, and the satellite is controlled by the scheduler.
Thus, if Starlink rolls out a FUP, it must be a significant update to the scheduler. If this update, for whatever reasons, includes bugs that cause it to fail, it may affect not just the FUP side, but the entire service, as we saw.
I would also argue that the issue was fixed by a rollback of the scheduler, and thus we won’t see FUPs being enforced yet.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
