How Tencent Cloud’s “Mi Master” Enables Scalable, Secure Mobile Payments

Background

Mobile payments have reshaped user experience and business models across many industries, and the market is now moving into a deep‑integration phase. Tencent Cloud’s “Mi Master” leverages over a decade of payment and billing experience from Tencent’s own services to offer a professional, one‑stop payment and billing solution for various sectors.

System Framework

To handle the massive transaction volume of products such as King of Glory, QQ Music, and WeChat Ads, “Mi Master” abstracts its infrastructure into three major subsystems:

Online Transaction Cluster : Provides a unified payment experience, standardizes transaction flow, and applies risk‑strategy management to ensure high reliability.

Offline Reconciliation Cluster : Performs real‑time and batch calculations on live streams to achieve full‑process accounting, precise user‑paying‑profile generation, and accurate settlement.

Operation Platform : Offers second‑level monitoring, one‑click scaling, and maintains system availability at 99.99%.

Real‑Time Transaction Assurance

Consistency is critical because any failure in the multi‑step payment‑to‑delivery chain invalidates the whole order. When a deduction request to a bank times out, “Mi Master” demotes the transaction from synchronous to asynchronous processing, handing control to an order center that monitors for eventual deduction results. If no result arrives, the order is further downgraded to offline processing, guaranteeing consistency even under channel‑deduction anomalies.

Additionally, the platform continuously synchronizes transaction data for matching, analysis, and alerting, enabling rapid detection of forged receipts or DB tampering through an audit system.

Traffic Scheduling and High Availability

During holidays or promotional events, request volume can surge to dozens of times the normal level. “Mi Master” employs both external and internal traffic‑scheduling mechanisms:

External Scheduling : A unified interface layer masks external differences and applies tiered flow‑control policies (caching, async handling, rate‑limiting) to prioritize transaction consistency while providing graceful user experience.

Internal Scheduling : Built on Tencent’s self‑developed TDF framework, a distributed transaction engine uses state machines and registration mechanisms to route traffic away from bottlenecks, ensuring high‑consistency execution across the system.

Physical isolation of SETs (service execution tiers) allows large‑scale traffic to be directed to dedicated SETs during peaks, with one‑click expansion to protect other business flows.

Scenario‑Based Stress Testing

To anticipate capacity limits, “Mi Master” applies a scenario‑driven load‑testing method directly on the production environment. It constructs pressure scenes that mimic the most demanding promotional activities—flash sales, coupons, red‑packet distributions, etc.—and gradually ramps up traffic during low‑peak periods. When a bottleneck is observed, testing stops and the system is scaled out, allowing capacity assessment and early problem detection.

Multi‑Dimensional Risk Control

The platform tackles black‑market threats through a three‑stage approach: pre‑detection, real‑time detection, and post‑audit. By aggregating diverse data sources, behavior statistics, and correlation analysis, it builds a “DNA” profile of malicious users and applies dynamic models with multi‑level strategies to block fraudulent transactions while minimizing false positives.

Conclusion

Through a combination of a three‑tier architecture, robust consistency mechanisms, flexible traffic scheduling, scenario‑based capacity testing, and comprehensive risk‑control, “Mi Master” delivers a secure, highly available payment ecosystem that can support large‑scale, real‑time mobile payment scenarios across industries.