What is a ledger database?
A ledger database is a database purpose-built to store financial transaction records with immutability, double-entry accounting constraints, and a complete audit trail. It is the technical infrastructure layer that a payment ledger runs on. Where a general-purpose database stores and updates data, a ledger database only appends. No record can be modified or deleted once written.
This matters because financial data has different integrity requirements than most other data. A corrupted transaction record is not just a bug. It is a lost fund, an incorrect balance, or a compliance violation.
How does a ledger database differ from a regular database?
Standard relational databases like PostgreSQL or MySQL can be used for financial record-keeping, and many early fintechs start this way.
The problem is that these databases do not enforce financial accuracy natively. Immutability, idempotency, and double-entry constraints have to be implemented in application code, where they can be bypassed by a bug or eroded as the codebase grows.
A purpose-built ledger database enforces these properties at the infrastructure level:
- The application cannot write a non-double-entry transaction
- No record can be overwritten or deleted
- Duplicate transactions are rejected by design
The correctness guarantee moves from application-layer convention to database-layer enforcement.
What are the core properties of a ledger database?
Here are the main properties
Immutability: The database is append-only. New transactions create new records. No existing record can be overwritten or deleted. This produces a tamper-evident history of every financial event, which is essential for audits and dispute resolution.
Idempotency: Duplicate transactions are prevented by design. If the same payment instruction is submitted twice due to a network retry or a bug, only one transaction is recorded. Most ledger databases implement this via idempotency keys on each write.
Double-entry bookkeeping: Every transaction credits one account and debits another by the same amount. Enforcing this at the database layer means no partial failure or race condition can produce an unbalanced entry.
Concurrency controls: A ledger database handles high volumes of simultaneous writes without allowing operations to interfere in ways that produce incorrect balances.
What is the difference between a ledger database, a payment ledger, and a general ledger?
These three terms are related but describe different things.
A ledger database is what you build on. A payment ledger is what you build. A general ledger is what the payment ledger feeds into.
Is a ledger database the same as a blockchain?
No. A blockchain is a specific type of distributed ledger database where records are organized in cryptographically linked blocks and validated by a decentralized network.
It shares the immutability property but adds decentralization, which most enterprise financial applications do not need. Most payment platforms use a centralized ledger database for their core financial records. Blockchains are used for on-chain settlement of crypto assets including stablecoins, where decentralization and public verifiability are the point.
Why do payment platforms use dedicated ledger databases?
As payment platforms scale, general-purpose databases create operational problems that ledger-specific infrastructure solves:
- Balance accuracy: Race conditions in a standard database can produce incorrect balances at high volumes. Ledger databases with proper concurrency controls prevent this
- Audit trails: Regulators and auditors expect a complete, tamper-evident transaction history. An append-only ledger makes reconstruction trivial
- Reconciliation: Double-entry enforcement at the database layer means every transaction is guaranteed to balance before it is written, which simplifies downstream reconciliation
- Virtual accounts: Platforms holding funds for multiple clients in pooled accounts use ledger databases to track each client's sub-balance with the accuracy required for pass-through FDIC insurance and regulatory compliance
Commercial options include Modern Treasury Ledgers, Twisp's Financial Ledger Database, and Tigerbeetle, an open-source high-performance financial database. Amazon QLDB was deprecated in 2025. Most larger fintechs build proprietary ledger systems on top of Postgres, accepting the tradeoff of custom enforcement over out-of-the-box guarantees.
