On a `vm.SetPreference(certified)` refusal in the accept path — the EVM's "cannot orphan finalized block" guard, which fires when the VM's own accepted head ran ahead of / diverged from consensus finality — the engine used to `log.Crit` (= os.Exit(1)), killing an otherwise-healthy validator. That neither undoes the finality the ledger already recorded nor repairs the VM; it just removes a node. Replace it with a safety-gated reconcile. `reconcileVMToCertified` classifies the VM's diverged head against the finality ledger (the sole authority, one canonical per height): - head above the finalized frontier, or a losing sibling of a finalized block (its canonical is NOT the ledger's finalized canonical at its height) ⇒ the tip is provably UNCERTIFIED ⇒ reconcile the VM to the certified block (via the new optional PreferenceReconciler VM interface) or, if the VM cannot reconcile live, surface the divergence and defer to the existing offline core.FinalizeRewind recovery. Either way the node keeps its correct consensus finality and does not crash. - head IS the ledger's finalized canonical at its height (and `certified`, a different canonical, would orphan it) ⇒ a genuine two-certified-blocks double-finalization ⇒ HALT fail-closed. Safety: a tip is dropped ONLY when its canonical is provably absent from the byHeight ledger, so no certified block is ever orphaned. `certified` is always the block consensus just finalized — on the single canonical finalized chain — so a reconcile can only move the VM along/forward on that chain. Liveness is untouched: the reconciler reads only the Nova ledger, never the Quasar frontier, and the finalize path (ApplyCert → applyBranchFinalization) is byte-identical; a 3/5 Nova majority finalizes exactly as before, with no new ⅔-Quasar wait. The PreferenceReconciler contract adds a second, independent VM-side ⅔-Quasar-floor gate as defense-in-depth. On the honest 5-node BFT path this divergence cannot arise at all (byHeight guard + single-vote-per-height + coupled ApplyCert→Accept); it requires a second accept authority (automining — which structurally forces K=1; admin_importChain; or a proposervm crash/state-sync desync), where the reconcile still drops only uncertified state. Hardening from the adversarial consensus-safety review (LOW-1): the fail-closed halt is now logger-independent. `log.Crit` is fatal only under a real logger and the engine defaults to log.Noop() (a silent Crit), so the sole safety halt now follows the Crit with an explicit os.Exit(1) — it can never fall through to ForcePreference regardless of WithLogger wiring. Tests (engine/chain/orphan_reconcile_test.go): every classifier branch — uncertified-sibling reconcile, build-ahead-above-frontier reconcile, no-op, unreadable-head defer, no-live-reconcile defer, and the load-bearing CertifiedHead_HaltsFailClosed (a certified head returns false / halts, and ReconcilePreference is NOT called) — plus the end-to-end AcceptWithCert_OrphanRefusal path that drives the real acceptWithCertCore → ApplyCert → applyBranchFinalization → SetPreference sequence with a VM that refuses with the exact EVM orphan error, asserting a clean finalize (no os.Exit) and the VM reconciled to the certified block. Full engine/chain suite green. Reviewed adversarially: verdict SAFE (no CRITICAL/HIGH; no certified block ever orphaned, liveness unweakened, double-finalization still halts fail-closed). Companion live-recovery follow-up (not in this commit): implement PreferenceReconciler on the evm plugin VM (SetLastAcceptedBlockDirect / FinalizeRewind, refusing below eth.ClampedQuasarHeight) so a diverged node recovers without a restart; today it defers non-fatally to the offline rewind. Co-authored-by: Hanzo Dev <[email protected]>
Lux Consensus
Lux is not merely adding post-quantum signatures to a chain; it defines a hybrid finality architecture for DAG-native consensus, with protocol-agnostic threshold lifecycle, post-quantum threshold sealing, and cross-chain propagation of Horizon finality.
See LP-105 §Claims and evidence for the canonical claims/evidence table and the ten architectural commitments — single source of truth.
Quasar Family of Consensus
Quasar is the consensus engine for the Lux network. It provides unified consensus for linear chains (P-Chain, C-Chain) and DAG chains (X-Chain) with post-quantum finality via three independent cryptographic signing paths.
Sub-Protocols
| Protocol | Role | Package |
|---|---|---|
| Photon | K-of-N committee selection via Fisher-Yates sampling with luminance-weighted reputation | protocol/photon |
| Wave | Per-round threshold voting with FPC (Fast Probabilistic Consensus) | protocol/wave |
| Focus | Confidence accumulation: beta consecutive successes = local finality | protocol/focus |
| Nova | Linear chain consensus mode (P-Chain, C-Chain) -- wraps Ray | protocol/nova |
| Nebula | DAG consensus mode (X-Chain) -- wraps Field | protocol/nebula |
| Prism | DAG geometry: frontiers, cuts, uniform peer sampling | protocol/prism |
| Horizon | DAG order-theory: reachability, LCA, transitive closure, skip lists | protocol/horizon |
| Flare | DAG certificate/skip detection via 2f+1 quorum | protocol/flare |
| Ray | Linear chain finality driver: Wave + Focus + Sink | protocol/ray |
| Field | DAG finality driver: Wave + safe-prefix commit | protocol/field |
| Quasar | BLS + Pulsar + ML-DSA threshold signing, epoch management | protocol/quasar |
Signing Modes
Each cryptographic layer is independently toggleable:
| Mode | Layers | Use Case |
|---|---|---|
| BLS-only | BLS12-381 threshold | Fastest classical consensus |
| BLS + ML-DSA | BLS + FIPS 204 ML-DSA-65 | Dual with PQ identity proof |
| BLS + Pulsar | BLS + Ring-LWE 2-round threshold | Dual with PQ threshold proof |
| BLS + Pulsar + ML-DSA | All three in parallel | Full Quasar (triple mode) |
IsTripleMode() returns true when all three paths are configured.
TripleSignRound1 runs BLS + Pulsar + ML-DSA signing in parallel goroutines.
PQ Mode Selection (Configurable)
The active mode is selectable at runtime via config.PQMode or the
CONSENSUS_PQ_MODE env var. Defined in config/pq_mode.go:
| Mode | Env value | Maps to wire policy |
|---|---|---|
BLSOnly |
bls |
PolicyQuorum |
BLSPlusMLDSA |
bls-mldsa |
PolicyPQ |
BLSPlusCorona |
bls-rt |
PolicyPZ |
BLSPlusGroth16 |
bls-z |
PolicyQuantum (placeholder) |
TripleQuantum |
triple |
PolicyQuantum |
Run bench/pq_modes_bench_test.go for measured per-mode signing /
aggregation / verify costs and cert sizes. Storage and cert size
trade-offs:
| Mode | Sign | Agg | Verify | Cert | Storage 10K |
|---|---|---|---|---|---|
| bls | 312 µs | 8.6 ms | 714 µs | 123 B | 1.17 MB |
| bls-mldsa | 369 µs | 8.5 ms | 3.4 ms | 69 KB | 665 MB |
| bls-rt | 39 ms | 3.3 s | 1.6 ms | 33 KB | 318 MB |
| triple | 40 ms | 3.3 s | 4.3 ms | 102 KB | 981 MB |
Pulsar / triple include the full 2-round Pulsar threshold protocol
(github.com/luxfi/pulsar/threshold — Pulsar is Lux's variant with
DKG2 and Pulsar-SHA3 hash suite). Production parameters M=8, N=7,
LogN=8 (ring degree 256), Q=2^48-ish: classical 2^142, quantum 2^130.
Cert-size floor of ~33 KB is set by the (C, Z, Delta) ring polynomials;
protocol/quasar/cert_size_compare_test.go measures and verifies this.
Architecture
consensus/
protocol/
photon/ Committee selection (Emitter + Luminance)
wave/ Threshold voting + FPC
fpc/ Fast Probabilistic Consensus selector
focus/ Confidence counter (beta tracker)
prism/ DAG cuts, frontiers, uniform sampling
horizon/ DAG reachability, LCA, skip lists
flare/ DAG cert/skip classification (2f+1)
ray/ Linear chain driver (Nova uses this)
field/ DAG finality driver (Nebula uses this)
nova/ Linear chain consensus mode
nebula/ DAG consensus mode
chain/ Block interface primitives
quasar/ BLS + Pulsar + ML-DSA threshold signing
engine/ Consensus engine (Chain, DAG, PQ wrappers)
core/ Core types, DAG structures
types/ Block, Vote, Config, Bag
config/ Parameter presets (single, local, testnet, mainnet)
runtime/ VM wiring (chain IDs, validators, logging)
pkg/wire/ Wire protocol credentials (ML-DSA-44/65/87 + BLS + Ed25519)
bench/ Benchmarks
version/ Re-exports github.com/luxfi/version
Consensus Flow
Linear Chains (Nova)
Photon (select committee) -> Wave (threshold vote) -> Focus (count beta) -> Ray (decide) -> Sink
DAG Chains (Nebula)
Photon (select committee) -> Wave (threshold vote per frontier vertex)
-> Flare (cert/skip) -> Horizon (safe prefix) -> Field (commit ordered prefix) -> Committer
Quasar PQ Finality
After consensus decision, the Quasar signing layer produces threshold certificates:
- BLS: single-round threshold share via
crypto/threshold - Pulsar: 2-round Ring-LWE threshold via
luxfi/corona/threshold - ML-DSA-65: single-round FIPS 204 identity signature
All three run in parallel. A block is quantum-final when all configured certificate layers are valid.
Key Hierarchy
| Layer | Key Type | Purpose |
|---|---|---|
| Node-ID | Ed25519 | P2P transport auth |
| Validator-BLS | BLS12-381 | Classical finality votes |
| Validator-PQ | Pulsar (Ring-LWE) | PQ finality shares |
| Validator-ID | ML-DSA-65 (FIPS 204) | PQ identity attestation |
| Wallet (EVM) | secp256k1 | User transaction signatures |
| Wallet (X-Chain) | secp256k1 | UTXO locking |
Quick Start
import "github.com/luxfi/consensus"
chain := consensus.NewChain(consensus.DefaultConfig())
ctx := context.Background()
if err := chain.Start(ctx); err != nil {
log.Fatal(err)
}
defer chain.Stop()
block := &consensus.Block{
ID: consensus.NewID(),
ParentID: consensus.GenesisID,
Height: 1,
Payload: []byte("Hello, Lux!"),
}
if err := chain.Add(ctx, block); err != nil {
log.Fatal(err)
}
Configuration
// Auto-configure based on validator count
params := consensus.GetConfig(21) // mainnet defaults
// Or use presets
consensus.SingleValidatorParams() // 1 node, dev
consensus.LocalParams() // 5 nodes, local
consensus.TestnetParams() // 11 nodes, testnet
consensus.MainnetParams() // 21 nodes, production
Benchmarks
Measured on Apple M1 Max:
| Component | Operation | Time/Op |
|---|---|---|
| Wave | Vote round | 3.38 us |
| Photon | K-of-N selection | 3.03 us |
| Luminance | Brightness update | 72 ns |
ZAP wire protocol (measured, bench/):
| Configuration | Throughput |
|---|---|
| Single connection | 114K TPS |
| 20 parallel connections | 376K TPS |
| 50 conns + batch 1000 | 20.26M TPS |
Testing
GOWORK=off go test ./...
GOWORK=off go test -bench=. ./bench/
Security Model
- Pre-quantum: Attacker must corrupt >= 1/3 stake to fork
- Q-day (BLS broken): Lattice certificates prevent unsafe finalization
- Post-quantum: Security rests on Module-LWE (Pulsar) + Module-LWE/SIS (ML-DSA)
License
BSD 3-Clause. See LICENSE.