Orbital Runtime / Orbit Scheduler
Kubernetes for
Earth + Orbit
Where should this workload run? When? At what fidelity? Orbit Scheduler makes placement decisions based on orbital position, ground station visibility, energy availability, and latency requirements.
Q2 2026
Q3 2026
2027
Global Coverage, Smart Placement
Watch scheduling decisions happen in real-time as satellites pass over ground stations.
The Problem
Kubernetes, Nomad, and cloud schedulers assume always-on networks, stable power, and predictable latency. These assumptions are invalid for orbital compute.
| Traditional Scheduler | Orbit Scheduler |
|---|---|
| Assumes always-on connectivity | Predicts ground station windows, plans handovers |
| Ignores power source | Schedules around solar peaks and eclipses |
| Static latency estimates | Models orbital geometry, ISL hops, propagation |
| Workloads fail on timeout | Pre-migrates state before windows close |
Capabilities
Orbit-Aware Placement
SGP4/SDP4 orbital propagation integrated directly into the scheduler. Knows where every node will be, when it will have connectivity, and what its energy state will be.
Energy-Aware Scheduling
Shift workloads to solar peaks, defer non-urgent work during eclipse, balance battery state across the constellation.
Latency-Aware Routing
Route inference requests based on current orbital geometry. Account for ground station visibility, ISL hops, and processing queues.
Predictive Handover
Pre-migrate state before communication windows close. No dropped connections, no lost work, no timeout failures.
How It Works
from rotastellar import OrbitScheduler
scheduler = OrbitScheduler(api_key="...")
job = scheduler.submit(
workload="llm-inference",
model="llama-70b",
latency_sla_ms=200,
priority="normal",
prefer_orbital=True # Prefer orbital when viable
)
print(f"Routed to: {job.placement}")
print(f"Expected latency: {job.latency_ms}ms")
print(f"Energy source: {job.energy_source}"){
"placement": "orbital_node_1",
"reason": "energy_surplus",
"orbital_state": {
"solar_angle": 67.2,
"power_available_w": 847,
"eclipse_in_min": 43
},
"latency": {
"ground_to_orbit_ms": 12,
"processing_ms": 156,
"orbit_to_ground_ms": 14,
"total_ms": 182
},
"fallback": "earth_us_west",
"energy_savings_pct": 43
}Ready to schedule across Earth and orbit?
Get early access to the Orbit Scheduler simulator.