Graph Database Applications in Financial Technology: Ant Group’s Practices and Future Outlook

The talk, presented by Ant Group senior technical expert Fu Zhisong, introduces the application of graph databases in the financial sector, covering Ant Group's exploration and insights on graph technology.

FinTech Architecture Evolution: Three stages are described – business digitization, data‑driven development, and data‑intelligence, each driving increasingly sophisticated data analysis and real‑time decision making.

Challenges in the Data‑Intelligence Era: Rapid growth in business types, massive data volume (over a trillion daily records), complex relational structures, and the need for real‑time decisions demand advanced storage and processing capabilities.

Technical Response – Storage Structure Evolution: Transition from traditional databases to KV stores, NoSQL, and finally graph storage to handle multi‑dimensional, time‑varying relationships, enabling high‑performance graph queries and analytics.

Technical Response – Computation Paradigm Evolution: Shift from offline and near‑real‑time computation to fully real‑time processing to support scenarios such as instant risk control and loan approval.

Ant Group’s Graph Storage Practice: Introduction of the GeaBase (Graph Exploration and Analytics Database), a large‑scale distributed real‑time graph database designed for high concurrency, low latency, strong consistency, and scalability, supporting queries via GQL, Gremlin, and various graph algorithms.

Application Scenarios: Social applications like Ant Forest, fraud detection (e.g., order‑scam), and credit‑card cash‑out prevention, all requiring millisecond‑level graph updates and queries.

Standardization Efforts: Participation in international (ISO/IEC Graph Query Language), national (Graph Database System Technical Requirements), and industry standards (graph database white papers) to promote unified graph technology.

Future Outlook: Anticipated growth in graph size, increased security requirements, faster decision making, and the vision of a single graph serving as the foundational layer for fintech.

Q&A Highlights: Definitions of temporal graphs, their use in real‑time risk control, GeaBase’s support for temporal graphs, and its optimizations for financial scenarios.