BFT from First Principles: Safety, Liveness, and Quorums

Monthly research note. Theme: Distributed Systems Under Failure.

TL;DR

A focused memo on BFT from First Principles: Safety, Liveness, and Quorums: define the model, state the properties, then design the system so those properties remain true under failure and adversaries.

Key takeaways

• Expose protocol state (epoch/term/commit index) as first-class telemetry.
• Mixed-version operation is the default; upgrades must preserve invariants.
• Treat membership changes and compaction as protocol events—not operational details.
• Make boundaries boring: validate inputs, cap costs, and be deterministic where needed.
• Treat retries, reordering, and partial failure as default conditions.

Why this matters

• State compaction and snapshots are where correctness goes to die quietly.
• Most protocol bugs hide in timeouts, retries, and membership changes.
• If your protocol isn’t testable under reordering, it isn’t deployable.
• Observability must explain protocol state, not just latency.

Key questions

• Which components require determinism for reproducibility?
• What is the unit of ordering (per key, per partition, global)?
• What is your reconfiguration model (joint consensus, epochs, leases)?
• What is the compaction story (snapshots, log truncation, state transfer)?
• Which safety property is non-negotiable (no double-commit, no forks, no split brain)?
• What is the failure model (crash, byzantine, partitions, reordering)?

Assumptions

• Workload is skewed: hot keys exist and dominate.
• Reconfigurations happen mid-incident (the worst time).
• Nodes restart with partial state unless you prove durability.
• Partitions happen at multiple layers (network, DNS, LB, service mesh).

Non-goals

• Relying on global time for ordering without strong synchronization assumptions.
• Pretending backpressure is an implementation detail.

Model & invariants

For quorum-based protocols, the intersection property is the backbone of safety:

Write down the safety property first. If it’s not written, it’s not implemented.

Make overload explicit: admission control is a protocol boundary.

Security properties

• Downgrade resistance: negotiation can’t silently weaken security posture.
• Least authority: privileges are scoped by purpose and time.
• Replay resistance: duplicated inputs do not change outcomes.
• Evidence: critical actions emit verifiable audit events.

Failure modes

• Recovery paths that only work when nothing is broken.
• Config drift that weakens security posture over time.
• Observability gaps during incidents (missing evidence).
• Resource exhaustion (CPU/bandwidth/storage) turning into correctness failures.

Design sketch

flowchart TD
  client["Client"] --> leader["Leader"]
  leader --> log["Replicated Log"]
  log --> snap["Snapshot"]
  snap --> recover["Recovery / Catch-up"]
  leader --> reconfig["Reconfiguration"]

Implementation notes

Make the state machine explicit; then make persistence and networking boring.

Operational invariants to monitor:
- leader_changes_per_minute
- commit_index_monotonic
- snapshot_install_failures
- quorum_acks_latency_p99
- rejected_requests_due_to_admission_control

Verification strategy

• Stress + skew tests: hot keys, slow disks, noisy neighbors.
• Model checking the smallest core (timeouts, election, reconfiguration).
• Upgrade tests: mixed versions and rolling deploy invariants.
• Linearizability checks for read/write APIs that claim it.
• Jepsen-style fault injection: partitions + reordering + client retries.

Operational notes

• Rehearse region failover and reconfiguration under load.
• Expose protocol state: term/epoch, leader, commit index, config version.
• Make client behavior part of the system: document retry semantics.
• Rate-limit retries and apply admission control before saturation.
• Treat compaction and snapshot install as first-class SLOs.

What to monitor

• Error budget burn + tail latency under load.
• Invariant violation rate (should be ~0).
• Admission-control / rate-limit rejections (by reason).
• Retry/timeout rates by endpoint and client cohort.
• Authz failures and policy denials (unexpected spikes).

Rollback plan

• Preserve evidence (configs, artifacts, audit logs) to reconstruct what changed.
• Define an explicit rollback trigger (metrics + thresholds).
• Prefer backward-compatible changes; avoid “flag day” upgrades.
• Keep dual-write / dual-verify windows where appropriate.
• Use canaries and staged rollout; stop early when signals degrade.

Evidence

• Designing Data-Intensive Applications (Kleppmann) (1) — The systems-engineering baseline for correctness, replication, and failure.
  • Evidence: Replication and consistency tradeoffs as engineering constraints; use as reference when naming guarantees.
• Jepsen (2) — Testing correctness under partitions and faults.
  • Evidence: Turn faults into test cases; prioritize partition and clock-skew scenarios that violate user-visible guarantees.

Open questions

• Which invariants are violated first under overload: latency, availability, or correctness?
• What is the worst-case recovery time after a leader + disk failure?
• How do you prevent “operator fixes” from changing safety properties?
• Where does your protocol assume synchrony without admitting it?

Checklist

• Safety properties stated as invariants.
• Assumptions listed and reviewed.
• Costs bounded (CPU/memory/bandwidth) under adversarial inputs.
• Telemetry captures correctness signals.
• Failure modes enumerated with mitigations.
• Rollback plan rehearsed and automated.

Further reading

• Time, Clocks, and the Ordering of Events (Lamport) — Causality, ordering, and why clocks are tricky.
• Jepsen — Testing correctness under partitions and faults.
• In Search of an Understandable Consensus Algorithm (Raft) — Consensus with explicit state machines and practical tradeoffs.
• Paxos Made Simple (Lamport) — Agreement basics and the invariants that matter.
• Designing Data-Intensive Applications (Kleppmann) — The systems-engineering baseline for correctness, replication, and failure.
• Site Reliability Engineering (Google) — Error budgets, incident response, and reliability as an engineering discipline.