How Wacai Built a Scalable FinTech Architecture: 6 Key Design Strategies

Developing software that satisfies users is challenging; architects must grasp diverse requirements, balance conflicts, and systematically address each need. Wacai architect Wang Fuqiang shares six key points from the company’s internet‑finance technical exploration.

1. System Layer Separation

Wacai’s ecosystem spans four dimensions: membership center, cash flow, risk control, and product center, plus settlement functions. Across all systems, they enforce strategic separation, employing SOA for distributed computing as the system scales.

Cash‑flow data differs from typical data, growing over time rather than per user, so Wacai partitions cash flow by user and timeline, choosing Scala for consistency. Middleware bottlenecks are mitigated by service‑oriented optimizations and data‑layer shielding.

Beyond technical boundaries, responsibilities are divided to leverage individual strengths, emphasizing front‑end/back‑end separation. Initially, asynchronous calls generated protective layers, but ultimately the front‑end fully owns its responsibilities, boosting overall efficiency.

2. Message Passing

After isolating layers, inter‑system communication is required. Wacai adopts a message‑passing mechanism, using remoting (distributed processing) and RPC concepts, with Dubbo as the service framework. For their financial workload, HTTP protocol suffices.

3. Asynchronous Processing

To minimize the impact of events, Wacai implements asynchronous handling wherever possible. When mobile requests reach millions, they close the loop with async processing, pushing results back after completion. Kafka and Akka (an actor model implementation for Java/Scala) are the primary tools.

Akka’s unrestricted throughput necessitates flow control; Wacai employs ACK mechanisms to alleviate backend pressure.

4. Information Storage – “More Is Better”

Wacai retains every change instead of overwriting, ensuring traceability. Kafka’s multi‑replica storage supports this approach. Core data changes now receive version numbers, enabling rollback and preventing data loss, effectively providing a “time‑machine” for troubleshooting.

5. System Security

Security is paramount in financial systems. Wacai meets ATBS baseline standards, establishing entry barriers to reduce technical costs. Defensive layers include network and application firewalls, while anti‑fraud measures leverage third‑party services (e.g., Tongdun) and manual risk‑control interventions.

Physical isolation of Kafka clusters prevents a single point of failure from affecting the entire system.

6. Storage Redundancy

Unlike typical internet services that prioritize availability (CP), fintech demands strong consistency. For core transaction systems, Wacai employs a simple, flexible multi‑write strategy to ensure data consistency and avoid message‑queue weaknesses.

The ultimate goal is a fully event‑driven, time‑and‑data‑oriented reactive system.