blockers: enumerate Tier C → A path (9 blockers: construction selection, academic basis, spec, ref impl, KAT, interop, proofs, package docs, cryptographer review)

2026-07-27 02:53:47 +00:00 · 2026-05-18 22:48:23 -07:00
parent 320a259cc0
commit 4e8d1657ed
1 changed files with 104 additions and 0 deletions
@@ -0,0 +1,104 @@
+# Magnetar — Blockers to NIST MPTC submission
+
+> Honest enumeration of what blocks Magnetar from moving Tier C →
+> Tier B → Tier A on the submission-readiness ladder. This is the
+> path-to-production register; see `SUBMISSION-STATUS.md` for the
+> tier definitions and phased plan.
+
+## Tier C → Tier B blockers (construction-level)
+
+### BLK-1 — Construction not selected
+
+No fixed threshold-SLH-DSA protocol. Multiple candidates exist
+in the literature:
+
+- **Komlo/Goldberg-style FROST-of-SLH-DSA** — would mirror Pulsar's
+  protocol shape but SLH-DSA's hash-tree traversal has no linear-
+  aggregation identity. Cannot produce byte-equal FIPS 205
+  signatures from threshold inputs.
+- **Full-MPC hash-computation** — run every SHAKE / SHA-2 inside
+  MPC. Slow (per-signature MPC ceremony dominates) but correct.
+- **Threshold seed-DKG + batched per-signature ceremony** —
+  hybrid: distribute seed via VSS, batch per-signature
+  reconstruction.
+
+Each has different security models, performance, and proof
+obligations. Selection requires research, not engineering.
+
+### BLK-2 — Reference paper / academic basis
+
+The threshold-SLH-DSA literature is sparse compared to threshold-
+Schnorr / threshold-Lattice. Magnetar's construction MUST cite a
+peer-reviewed paper as its academic basis (analogous to Pulsar's
+ePrint 2024/1113 for the Corona-family). Without this, the
+construction risks being unreviewed work.
+
+### BLK-3 — Spec definition
+
+No `SPEC.md` exists. Spec authoring is blocked on BLK-1
+(construction selection).
+
+## Tier B → Tier A blockers (engineering-level)
+
+These activate once BLK-1, BLK-2, BLK-3 close.
+
+### BLK-4 — Reference implementation
+
+No Go implementation. Layout: `ref/go/pkg/magnetar/` mirroring
+Pulsar. Estimated 4-8 weeks once construction frozen.
+
+### BLK-5 — KAT vectors
+
+Deterministic KAT generator + replay tests. Cannot exist without
+implementation.
+
+### BLK-6 — Cross-validation harness
+
+Third-party FIPS 205 verifier integration (cloudflare/circl, pq-
+crystals reference) for the threshold-output-verifies-under-
+unmodified-FIPS-205 claim (analog of Pulsar's Class N1 evidence).
+
+### BLK-7 — Proof artifacts
+
+- EasyCrypt theories for threshold-SLH-DSA correctness
+- Lean bridges for the algebraic identities (TBD — SLH-DSA is hash-
+  based, the bridge story is different from lattice schemes)
+- Jasmin constant-time analysis of the per-signature MPC ceremony
+
+Hash-based constructions have well-developed CT analysis pipelines
+(libjade SLH-DSA exists), so the single-party CT side is upstream-
+ready. The threshold MPC overlay is novel work.
+
+### BLK-8 — Submission package documentation
+
+16-doc mirror of Pulsar's structure adapted to Magnetar specifics.
+Blocked on BLK-1 + BLK-4.
+
+### BLK-9 — Independent cryptographer review
+
+Independent reviewer must attest construction + impl + proofs +
+tests. Blocked on BLK-1 through BLK-8.
+
+## Non-blockers
+
+These are NOT blockers — they are deliberate non-claims that
+Magnetar will continue to make even at Tier A:
+
+- **Byte-equal to single-party FIPS 205**: SLH-DSA's hash-tree
+  traversal precludes this. Magnetar will instead claim
+  "verifies under unmodified FIPS 205 given seed-DKG soundness"
+  (a strictly weaker but still well-defined claim).
+- **Sub-second per-signature**: per-signature MPC ceremony is
+  hash-dominated; production deployment will batch / amortize but
+  Magnetar will not pretend to match Pulsar's signing throughput.
+- **Production-ready before 2027**: see `SUBMISSION-STATUS.md`
+  target window.
+
+## Cross-references
+
+- `README.md` — repo purpose + status
+- `DESIGN.md` — research direction
+- `SUBMISSION-STATUS.md` — tier definitions + phased plan
+- [`luxfi/pulsar`](https://github.com/luxfi/pulsar) — Tier A template
+- [`luxfi/quasar`](https://github.com/luxfi/quasar) — Polaris profile
+  cannot deploy until Magnetar matures to Tier A