In the rapidly evolving internet entertainment industry, iQIYI, prioritizes user needs and drives innovation to create an integrated “long-form + short-form” content ecosystem. This ecosystem spans long-form dramas, micro-dramas, micro-variety shows, animations, films, games, novels, IP derivatives, and offline entertainment—pioneering diversified business models for video platforms.

iQIYI’s customer loyalty points system is a critical backbone for user engagement, membership rewards, and cross-platform synergy, playing an essential role in driving business growth. However, as the business expanded, its loyalty points system faced mounting technical challenges.

Navigating the challenges of a dual-database architecture

iQIYI’s current loyalty points system serves four major business units: Lite Edition, Baseline Services, International Services, and Integrated Terminal Services. The system needs to manage data across various loyalty points schemes and ensure seamless handling of diverse scenarios within each business unit.

In some cases, a single user has multiple points accounts—each tied to a different points line —adding complexity and demanding both high scalability and massive concurrency.

The dual-storage architecture included:

• A MySQL database that stored aggregated points values, managing operations such as balance adjustments, total value queries, transaction logs, and counters.
• A MongoDB Atlas database that stores detailed point transaction histories, supporting historical queries and audits.

While this setup worked initially, exponential growth in users and transactions began exposing critical limitations:

• High-concurrency write bottlenecks: MySQL instances struggled to scale horizontally fast enough during peak loads such as major promotional events, hitting write limits and causing significant performance degradation.
• Data consistency challenges: Separation of aggregated points in MySQL and transaction histories in MongoDB required complex distributed transactions or dual-write compensation mechanisms, increasing operational risk.
• Operational overhead from a hybrid architecture: Running two different database systems created fragmentation, making scaling, monitoring, and maintenance cumbersome. Developers had to be proficient in both technologies, which raised costs and complexity.

Adopting MongoDB’s unified storage architecture

To streamline its architecture, improve scalability, and eliminate dual-database complexity, iQIYI’s technical team evaluated various database solutions. It ultimately migrated the points system entirely to ApsaraDB for MongoDB 7.0. The move consolidated both aggregated point totals and transactional details into a unified platform, unlocking four core benefits:

• Enhanced scalability and concurrency: MongoDB 7.0’s native sharding enables horizontal scaling across hundreds of millions of users and multiple business lines. During traffic spikes, such as large promotional events, data automatically distributes across multiple nodes. This reduces system pressures and maintains high availability and stability.
• Robust transactional consistency: The points systems require strict consistency for operations, especially for top-ups and redemptions, to ensure synchronized totals and transaction histories. MongoDB 7.0 supports robust multi-document transactions and configurable write concerns, enabling reliable synchronization and rollback capabilities during transactional workflows.
• Flexible data modeling for fast business evolution: Each business unit often requires unique point rules and entitlement logic. MongoDB’s document model supports evolving schemas without downtime, facilitating the addition of new business rules and fields without disrupting operations.
• Simplified operations and maintenance: ApsaraDB for MongoDB offers built-in monitoring, alerting, and automation. Combined with managed cloud services, this reduces infrastructure costs and minimizes manual maintenance, allowing developers to focus on building applications that support the business, rather than managing database operations.

Migrating the loyalty points system to ApsaraDB for MongoDB marked a strategic shift for iQIYI—from using MongoDB as supplementary storage to making it a core platform for high-performance, high—concurrency workloads.

This migration resolved critical performance bottlenecks, ensured transactional consistency, and created a scalable, flexible foundation for future growth. By transforming its points system from a basic rewards tool into a powerful data engine, iQIYI is driving smarter, scalable operations. and seamless user experiences across its ecosystem. The company has now set a benchmark for intelligent, data-driven loyalty and rewards platforms across industries.