Stripe has developed a robust system to handle the complexities of global payment processing, ensuring high availability and reliability. The foundation of this system is the Global Payments and Treasury Network (GPTN), which supports over 135 currencies and payment methods through partnerships with banks and financial networks in 185 countries. To manage the diverse interfaces, data models, and behaviors of these entities, Stripe developed Ledger, an immutable and auditable log of financial data.

Ledger serves as the system of record for Stripe’s financial transactions, providing a standardized representation of money movement. It supports the automated Data Quality (DQ) Platform, ensuring that Stripe accurately manages money for users and meets the needs of both internal customers and external auditors. The Ledger system processes five billion events daily, with 99.99% of dollar volume ingested and verified within four days.

The design of Ledger involves modeling internal data-producing systems with common patterns and proactive alerting to surface issues. The system encodes a state machine representation of producer systems, tracking the movement of balances between accounts. It ensures data accuracy and consistency by computing account balances and generating comprehensive statistics.

Ledger abstracts the complexities of different systems by representing all activity on the Stripe platform as a common data structure. This allows for unified monitoring and analysis, providing mathematical proof of correctness through double-entry bookkeeping. This approach guarantees that all money in the system is fully accounted for by balancing credits and debits.

The DQ Platform built on Ledger unifies the detection of money movement issues and response tooling. It measures the health of fund flows through metrics such as clearing, timeliness, and completeness. These metrics ensure that internal stakeholders can make sound judgments about the correctness of underlying data systems. The platform maintained a 99.999% readiness target, even as data volume grew tenfold.

Clearing measures the fraction of Ledger that is appropriately zeroed out at steady state, ensuring no unresolved balances. Timeliness guarantees that data arrives on time for time-sensitive functions like monthly report generation. Completeness ensures that all data from upstream systems is accounted for, using cross-system checks and automated anomaly detection.

To address the variations in data systems, Ledger represents the movement of balances between discrete states, turning the process into an observable state machine. This abstraction allows for monitoring handoffs between data pipelines and ensures consistency between disconnected systems. The system tracks individual transactions through their entire lifecycle, providing detailed tracing and analysis.

When issues arise, Ledger’s immutability allows for accurate auditing and analysis by maintaining a complete log of all events. This ensures that past states can be reconstructed and examined for discrepancies. Data correction involves reverting and reprocessing prior operations to maintain the integrity of Ledger as the system of record.

The hierarchical automated alerting and rich tooling built on the DQ Platform provide interactive metric displays, analysis, and guidance for internal teams. This system allows for proactive feedback, simple data manipulation, and meaningful metrics tailored to specific business use cases. Leaders can view team-level DQ metrics, including the financial impact of issues, and take appropriate actions to resolve them.

In conclusion, Stripe’s Ledger and DQ Platform provide a robust framework for managing the complexities of global payment processing. By standardizing data representation and implementing proactive monitoring and alerting, Stripe ensures high accuracy, consistency, and reliability in its financial operations. This system supports Stripe’s growth and ability to handle increasing data volumes while maintaining stringent quality standards.

Learn More