Customer Introduction
Zhuhai Sandbox Network Technology Co., Ltd. was founded in 2015, positioning itself as a global game user generated content (UGC) platform dedicated to helping everyday players turn their creative ideas into games.
Blockman Go is a 3D pixel-style multiplayer online sandbox game combining building, shooting, survival, and social gameplay. Built around an open world where players can freely construct or destroy terrain, the game features strong social features, including customizable characters, clan building, and a built-in map editor, Blockman Go has cultivated an active UGC ecosystem across Europe, Asia-Pacific, North America, and South America.
Challenges & Goals
As Blockman Go scaled globally, Sandbox Network faced four key operational challenges:
After years of live operations, the game has accumulated tens of millions of player records, including character progression, in-game assets, and social relationships, with individual tables potentially reaching hundreds of millions of rows, pushing the traditional MySQL system to its storage and query performance limits.
With millions of players online simultaneously, core scenarios such as combat state synchronization and item trading demanded millisecond-level response times, requiring the database to sustain over 100,000 QPS.
The platform's continuously evolving mini-games brought frequent changes to data structures, making traditional SQL schema modifications costly and inefficient. This resulted in increasing development overhead and operational risk. In addition, the semi-structured data created relational schemas that struggled to model and scale effectively.
Why Tencent Cloud
Tencent Cloud's deep expertise in database architecture for large-scale gaming platforms made it the ideal partner to address Blockman Go's data scalability, concurrency, and UGC flexibility challenges.
TDSQLC’s computestorage separation and distributed architecture support PB scale data, automated scaling, and second level snapshot recovery, enabling fast rollback and high data reliability without manual sharding.
TencentDB for Redis stores high frequency, real time data including player status, global leaderboards, and item inventory. By leveraging a sharded cluster architecture, it supports million level concurrent access with ultra low latency read/write performance and stable throughput, meeting the stringent demands of real time gameplay and global ranking systems.
MongoDB’s BSON document model supports nested JSON and dynamic fields without schema changes, enabling fast configuration updates. Independent compute and storage scaling allows a single cluster to handle up to 128 TB of data, supporting continuous content growth.
Highlighted Benefits
1.Scalable storage and high-concurrency processing
TDSQL-C's compute-storage separation architecture breaks through one-node capacity limits, supports billion-level player data tables and complex queries with ease. In parallel, clustered Redis architecture with sharding maintains stable, high-performance caching under massive concurrency.
2.Flexible data modelling for rapid UGC iteration
MongoDB's document-based data model stores heterogeneous game configurations natively, allowing new item attributes, skill rules and nested logic to be added without schema changes, removing a key constraint for the platform's evolving UGC mini-games.
3.Automated Scaling and optimised operations
TDSQLC automates sharding and routing, while Redis and MongoDB clusters enable minute level horizontal scaling, eliminating manual database splitting and data migration, and significantly reducing operational complexity as the business scales.