Add static and stochastic Docker test harnesses

Move examples/docker-network/ to testing/static/ and add testing/chaos/
as a new stochastic simulation harness.

testing/static/ — Static 5-node test harness:
- Fixed mesh, chain, and mesh-public topologies with docker compose
- Manual test scripts (ping, iperf, netem)
- Build script, config generation, key derivation

testing/chaos/ — Stochastic network simulation:
- Python orchestrator generating N-node FIPS meshes with dynamic
  network conditions, driven by reproducible YAML scenarios
- Topology generation: random geometric, Erdos-Renyi, or chain
  graphs with BFS connectivity guarantee
- Per-link netem: HTB classful qdiscs with u32 filters for per-peer
  impairment (delay, loss, jitter), stochastic mutation across
  configurable policy profiles
- Per-link bandwidth pacing: HTB rate limiting with configurable
  tiers (1/10/100/1000 mbps) randomly assigned per edge
- Link flaps: tc netem 100% loss with graph connectivity protection
- Node churn: docker stop/start with netem re-application on restart,
  shared down_nodes tracking across all managers
- Traffic generation: random iperf3 sessions between node pairs
- Down-node guards: all docker exec callers check container liveness,
  auto-detect crashed containers via is_container_running() safety net
- Log collection and post-run analysis (panics, errors, sessions,
  MMP metrics, tree reconvergence)
- chaos.sh wrapper with --seed, --duration, --verbose, --list options
- Four scenarios: smoke-10, chaos-10, churn-10, churn-20
This commit is contained in:
Johnathan Corgan
2026-02-20 13:35:57 +00:00
parent 12db6f561a
commit 66c268a564
46 changed files with 2827 additions and 71 deletions
+22
View File
@@ -0,0 +1,22 @@
# FIPS Testing
Integration and simulation test harnesses for FIPS, using Docker
containers running the full protocol stack.
## Test Harnesses
### [static/](static/) -- Static Docker Network
Fixed topologies (mesh, chain) with 5 nodes. Manual scripts for
building, config generation, connectivity tests (ping, iperf), and
network impairment (netem). Useful for deterministic debugging and
validating specific topology configurations.
### [chaos/](chaos/) -- Stochastic Simulation
Automated randomized testing with configurable node counts, topology
algorithms (random geometric, Erdos-Renyi, chain), and fault
injection (netem mutation, link flaps, traffic generation, node
churn). Scenarios are defined in YAML and executed via a Python
harness that manages the full lifecycle: topology generation, Docker
orchestration, fault scheduling, log collection, and analysis.