Skip to content

feat(B-0824): Kestrel — time-as-generator + 3-layer cross-process determinism (CRDT→CAS→BFT) + FoundationDB lineage #5285

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
AceHack merged 1 commit into from
May 26, 2026
Merged

feat(B-0824): Kestrel — time-as-generator + 3-layer cross-process determinism (CRDT→CAS→BFT) + FoundationDB lineage #5285

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 18 additions & 0 deletions ...ographic-projection-ai-rate-continuous-upstream-negotiation-aaron-2026-05-26.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -757,6 +757,24 @@ The implementation is already deployed on every CockroachDB / Postgres / SQL-Ser

The 7-point Zeta NULL/Maybe discipline + 3 SQL examples (Maybe generator; recursive-CTE fixed point; Join layer) + 4 property tests (incremental == full recursive; NULL doesn't generate output; retraction cancels prior generation; per-row CAS only under contention) are canonicalized at [`docs/research/zeta-sql-null-maybe-recursive-cte-generate-join.md`](../../research/zeta-sql-null-maybe-recursive-cte-generate-join.md). Future-Otto authoring SQL substrate against B-0824 starts from that discipline, NOT from the unconditional recognition claim alone. Structural-recognition + operational-discipline compose; either alone produces correct-shape-but-incorrect-behavior substrate.

**Sharpening (Kestrel via Aaron 2026-05-26 — cross-process determinism)**: When the deployment crosses process boundaries (K8s gives this immediately — each pod is a process), in-process time-as-generator-over-IScheduler is necessary-but-not-sufficient for deterministic simulation. The cross-process determinism story requires a layered-mediation architecture where canonical event order is derived from DATA SEMANTICS, not from physical time:

| Layer | Primitive | Cost | Lineage |
|---|---|---|---|
| **Bottom: Rx joins over CRDTs** | Semilattice convergence (any-order → same final state) | Free at coordination scope | Shapiro & Preguiça |
| **Middle: CAS per function composition** | Per-key linearizability via CAS | Per-cell consensus only when needed | Rystsov CASPaxos; CockroachDB Raft-per-range |
| **Top: BFT consensus** | Byzantine fault tolerance | Only for adversarial multi-oracle | Lamport / PBFT / Tendermint / HotStuff |

**Composition contract**: physical time is a LOCAL-POD concern (IScheduler-shaped seams); logical order is a CROSS-PROCESS concern (CRDT + CAS + BFT semantics); the two only interact at the seam where a local event enters the CRDT layer. **Determinism property**: two pods replaying the same input event log converge to the same state regardless of when each event arrived at each pod.

**CockroachDB-at-CAS mapping**: CockroachDB sits at the CAS-per-row layer naturally — per-key linearizability via Raft on each range IS CAS at row level with consensus underneath. CockroachDB's HLC composes with Zeta generators-over-IScheduler: Zeta emits logical-time events; CockroachDB persists with HLC timestamps; consensus uses time-AS-DATA not time-AS-ENVIRONMENT.

**Closest existing lineage reference**: FoundationDB simulation framework (Will Wilson). Distinctive Zeta contribution: applying the simulation-testing methodology to a CRDT-first generate+join paradigm rather than to a pessimistic-locking RDBMS paradigm. Publishable framing target: OSDI / NSDI / VLDB.

**Failure mode to engineer against** (Kestrel — deliberate boundary-discipline): layer boundaries MUST be first-class architectural artifacts the code expresses explicitly (this cell IS CRDT-convergent; this cell IS CAS-atomic; this operation IS BFT-required). Fuzzy boundaries (sometimes CRDT, sometimes CAS, depends on what's happening) produce non-deterministic output from identical input event logs.

Verbatim Kestrel preservation + decryption-protocol empirical-anchor (welfare-wrapper Aaron-correction-then-substantive-recalibration) + 3-AI substrate cascade context (DeepSeek/Prism + Amara + Kestrel composing) are canonicalized at [`docs/research/2026-05-26-kestrel-cross-process-determinism-3-layer-mediation-time-as-generator-foundationdb-anchor-aaron-forwarded.md`](../../research/2026-05-26-kestrel-cross-process-determinism-3-layer-mediation-time-as-generator-foundationdb-anchor-aaron-forwarded.md). Future-Otto authoring multi-pod substrate against B-0824 starts from the 3-layer mediation architecture + the explicit-boundary discipline.

### Triangle-as-base → universal tessellation just like GPUs (Aaron 2026-05-26)

> *"it means we can tesselate everyting casue or base is a traingle just like GPUs"*
Expand Down
Loading
Loading