Simulations and scenarios
Scenarios let you ask "what if" against the fleet without touching it: lose a gateway, add a constellation, put a candidate station on a mountain, soak a region in rain. Every scenario engine in the console is deterministic and pure: it mutates a copy of the scene snapshot, reports its results as modeled, and resets cleanly. The live fleet is never edited by a hypothesis.
A gateway outage scenario applied: the disabled gateway dimmed, re-homed feeder links drawn to their new gateways, orphaned satellites highlighted, and the before/after statistics card open.
Gateway outage
The flagship what-if: disable or remove one or more gateways (selected by id, country, or region) and see what the network does. The engine re-derives fronthaul access geometrically from the surviving gateways, using the same feasibility rules as the live scene, then reports:
- Re-homed satellites: each satellite that lost its serving gateway but found another, with its from and to gateways named.
- Orphaned satellites: satellites left with no feasible gateway at all.
- Before and after fleet statistics: the same deterministic metrics (utilization, capacity, contact) computed on both snapshots at the same sim time, so the delta is an honest subtraction, not an estimate.
Results are explicitly modeled: access links are re-derived from geometry, so they can differ from the replayed store's links, and the answer says so. An outage that matches zero gateways reports the empty match plainly rather than inventing an effect.
Synthetic orbits and stations
Design-side scenarios inject temporary assets rather than mutating existing ones:
- Orbit scenarios synthesize satellites or whole Walker constellations from a mission intent (altitude, orbit class, count, planes, inclination). The orbit is emitted as a standard TLE and injected through the exact pipeline uploads use, so the synthetic satellite renders, links, and scores identically to a real one; the only difference is its permanent SYNTHETIC provenance. Coverage summaries (combined and per-satellite coverage, pass intervals) are computed from the same element set the rendered asset uses, so the numbers match the picture exactly.
- Ground station scenarios place a labeled synthetic station and score the current fleet's contact utility against it: contact percent, contact minutes, passes, longest idle gap. The station is emitted as the same tabular content a real upload produces and qualifies as a Ka fronthaul gateway by the same rules. Satellites without propagable orbital elements are excluded from the evaluation and counted as excluded; visibility is never fabricated.
Both paths live behind plain requests to the assistant ("create a 24-satellite Walker at 550 km", "what would a station at Longyearbyen buy us"), and both leave inspectable assets on the globe under the Synthetic group in Assets.
Deterministic rain
There is exactly one rain model in the console, an ITU-R P.618 approximation inside the Ka fronthaul link budget, and no random weather anywhere. A weather scenario is an explicit field:
- Uniform rain: one rain rate (mm/h) at every station.
- Regional rain: rain cells defined as lat/lon boxes with a rate each; a station takes the maximum rate of the cells covering it, and zero outside all cells.
Rain resolves per station because attenuation happens on the low-elevation slant path above the site. Coverage and siting evaluations gate each sample on whether the link budget still closes at that sample's elevation and slant range under the scenario rain rate.
The reporting rule is fixed: clear-sky and rain-degraded numbers are always presented together. A site comparison under 25 mm/h rain shows what each site loses relative to clear sky, never just the degraded figure alone, so rain sensitivity is visible as a delta.
Saved scenarios
A scenario worth arguing about is worth reproducing. Saved scenarios persist in the browser and carry everything needed to rebuild the exact run:
- The exact synthetic injection payloads: the literal TLE text and ground-station table content that were injected, not a re-derivation.
- The sim epoch the scenario was evaluated at; loading a scenario seeks the timeline there.
- The what-if mutations (outages and their selectors).
- The evaluation config: window hours, step seconds, elevation mask, and the weather scenario.
Because the evaluation engines are deterministic (fixed step, no randomness, stable ordering), re-running a restored scenario at its saved epoch yields identical numbers. That is the reproducibility contract: a coverage percent quoted in a review can be regenerated to the digit weeks later.
Reset
Scenarios reset cleanly and completely: mutations were applied to copies, so reset restores the unmodified scene, and synthetic assets can be removed or kept as you choose. Multi-run studies (trade studies, design comparisons) reset between runs so each configuration scores independently on identical windows and masks.
Common workflows
| Task | How |
|---|---|
| Rehearse a site outage | "simulate losing the Oslo gateway"; read re-homed vs orphaned |
| Stress a region | Disable all gateways in a country, compare before/after statistics |
| Size a design | Create orbit scenarios per configuration, reset between runs, compare on one metric |
| Test rain resilience | Re-run a site comparison with a representative rain rate |
| Preserve a study | Save the scenario; reload it to reproduce the exact numbers |
| Clean up | Reset the scenario, delete or hide synthetic assets |
Example scenarios
The resilience question. Before a maintenance window, an operator simulates disabling both northern gateways at the planned time. Two satellites re-home to a mid-latitude site with tighter geometry; one is orphaned for 22 minutes of the window. The before/after statistics quantify the capacity dip, and the maintenance plan shifts by an orbit to close the orphan gap.
A design review that repeats. An engineering team saves the scenario behind their coverage claim: three injected TLEs, one candidate station, a 12-hour window at 60-second steps with a 10-degree mask, clear sky. In the follow-up review a month later, loading the scenario seeks the timeline to the saved epoch and reproduces every number in the original deck.