Skip to content

EX09 — Large LEO Service Network

Tags: showcase scale · communications · ISL routing · gateways Input: large-leo-service-network.yaml · Runtime tier: showcase-scale · Recommended views: 2D, 3D, Analysis

Purpose

Exercise regional demand, gateway delivery, and sparse ISL routing across a 1,584-satellite LEO shell for a full day.

Setup

The architecture uses 72 planes with 22 satellites each, two gateways, three service regions, a two-minute reporting cadence, and Walker-adjacent ISLs.

astraeus run astraeus-resources/examples/user/large-leo-service-network.yaml

The run is designed for a qualified workstation or VPS. Review the measured resource envelope before execution on the showcase page.

RunSpec Focus

The scheduler selects the lowest-latency visible gateway, permits up to 12 ISL hops, and exposes store-forward routing across a 12-step horizon.

transport:
  scheduler:
    path_mode: isl_gateway_routed
    gateway_selection: lowest_latency_visible
  isl:
    enabled: true
    topology: walker_adjacent
    max_hops: 12
    store_forward_horizon_steps: 12
    capacity_mbps: 300

Results

The reference run completed successfully and delivered all admitted bytes. Of the selected paths, 82.06% used more than one hop; routes averaged 2.54 hops, with p95 of five and a maximum of seven. The measured runtime, memory, output size, host context, and current preflight comparison are kept together on the large-scale showcase page.

Network result Value
Byte delivery rate 100%
Multihop route fraction 82.06%
Mean hops 2.54
p95 hops 5
Maximum hops 7

Artifacts To Inspect

Artifact What it shows
transport/isl_summary.json Sparse topology, routing closure, hop distributions, and memory diagnostics
transport/isl_routes.parquet Materialized selected paths and route-level latency components
transport/kpi_summary_gateways.parquet Delivered traffic and loading for Madrid and Honolulu
transport/transport_summary.json Fleet-level delivery, capacity, link margins, and scheduler outcomes

Interpretation

At this scale, sparse adjacency and bounded routing keep the full-day study tractable while retaining genuinely multihop behavior. The perfect delivery rate belongs to this offered-load and capacity configuration; hop distributions, gateway loading, and waiting latency reveal how the network achieved it.