Merkle Checkpointing

Cryptographic verification of audit log integrity.

How it works

Event hashing

Each event is hashed:

H(event) = SHA3-256(canonical_json(event))

Merkle tree construction

Events are organized into a Merkle tree:

        Root Hash
       /         \
    H(1,2)      H(3,4)
    /    \      /    \
  H(1)  H(2)  H(3)  H(4)

Checkpoints

Periodic checkpoints capture:

• Merkle root hash
• Timestamp
• Event count
• Previous checkpoint hash

Checkpoint structure

{
  "checkpointId": "uuid",
  "timestamp": "2024-01-15T00:00:00Z",
  "merkleRoot": "sha3-256-hash",
  "eventCount": 10000,
  "firstEventId": "uuid",
  "lastEventId": "uuid",
  "previousCheckpoint": "uuid",
  "signature": "dilithium-signature"
}

Verification

Verify checkpoint

GET /audit/v1/checkpoints/{checkpointId}/verify

Verify event inclusion

GET /audit/v1/events/{eventId}/proof

Returns Merkle proof:

{
  "eventId": "uuid",
  "eventHash": "hash",
  "proof": ["hash1", "hash2", "hash3"],
  "checkpointId": "uuid",
  "verified": true
}

Checkpoint frequency

• Default: Every 1 hour
• Configurable per tenant
• Minimum: 15 minutes
• Maximum: 24 hours

External anchoring

Checkpoint hashes can be anchored to:

• Public blockchain
• Trusted timestamping service
• Customer's audit system