Skip to content

Overview

Centurion topology optimization compares candidate validator deployments and selects the one that maximises risk-adjusted net value while surviving a declared set of failure scenarios. Two models are documented here.

Recommended model: Model 2 (Phase-1)

The winning topology is Model 2: 48 validators, evenly spread as 8 across each of six failure domains.

  • Profile: recommended_safe
  • Validators: 48 (3× AWS combined + 2× GCP combined + HOME VC-only, 8 each)
  • Protected scenarios survived: 52 / 52
  • Risk-adjusted annual net value: 86.52 CTN
  • Principal required: 1,536 CTN · Solver: scipy-highs (Gurobi-first formulation)

Every validator client subscribes to all five AWS/GCP beacon endpoints for cross-provider failover. 48 is the largest count that still survives every protected double failure: a 49th validator would drop a protected pair below the hard 2/3 finality floor.

→ Model 2 overview · diagram · final report · reports

Model 1 — Day-0 baseline (for context)

The earlier, AWS-anchored Day-0 launch plan. It is intentionally minimal and honest about its own limits (it is over the configured AWS budget and cannot survive a total-AWS outage); Model 2 supersedes it as the recommended Phase-1 topology.

  • Family: aws_core_plus_edge_plus_public_boot_sentry
  • V=18, T=12, c=4, e=3 · AWS Regions us-east-1, us-east-2, us-west-2
  • AWS CORE year-1 cost: $3,381.00 · Budget fit: no

→ Model 1 overview · diagram · reports

Regenerate & validate

python tools/optimize_topology_gurobi.py --reports-dir reports
python tests/test_topology_solution.py