The article should cover info about (if available?) the scalability of the system:
- Link Capacity between satellites (must be extremely high) - and inter-satellite peering will change frequently and cause interrupts?
- Routing capacity for each satellite - end-to-end path will be quite dynamic (routing in cabled networks is quite static once the first packet is routed and forwarding tables are established in routers end-to-end) - not so for routers in Starlink. Routing capacity must be extremely high?
- Link Capacity between ground nodes interfacing T1-T3 providers and satellites (must be huge compared to client dish up/down capacity) - is 10s of thousands of hub nodes planned or high capacity nodes?
- Not easy to upgrade HW in satellites to remove bottlenecks - must live with routing & laser capacity until replaced (satellite lifetime is limited and will over time be replaced by new satellites).