TDStore Engine
Last updated: 2025-07-24 10:19:51
Note:
Historical Archive
In industries such as e-commerce, retail, finance, O2O, and social applications, large amounts of historical data, together with the latest data, are stored in MySQL databases over time. The mixed storage of cold and hot data leads to high storage costs, decreased database performance, and backup difficulties. The TDStore engine features a massive storage capability, low cost, high reliability, and strong scalability, providing users with a large-capacity data archiving and storage scheme at very low costs. It also supports transaction processing and online data access. The TDSQL TDStore engine has been used by Tencent's communication top-up service and credit card repayment records.
Massive Data
Traditional databases face many challenges when storing and managing massive data, especially in high-concurrency read-write and high-throughput scenarios. In scenarios such as log streams, IoT, and smart homes, common business requirements include large data volumes, numerous single-table records, high write throughput, and high consistency. The TDStore engine with excellent horizontal scalability can increase the system's read-write throughput and processing capacity virtually linearly by adding nodes. It supports processing hundreds of TB or more of data. Additionally, the engine incorporates the Multi-Raft protocol, ensuring strong consistency and high availability. It also has fast online DDL and auto scaling capabilities, allowing users to store and manage massive data stably and efficiently. Furthermore, the TDStore engine provides efficient data compression and storage policies, effectively reducing storage costs. Tencent's invoice flow service has already adopted the TDSQL TDStore engine.
Agile Business
As modern market demands become increasingly variable and business R&D becomes more agile, flexible schemas and unrestricted auto scalability are particularly important for users in industries such as games and SaaS. The TDStore engine provides strong support for handling agile business needs. Its powerful auto scaling capability enables it to rapidly scale out or in resources when needed. It supports both horizontal and vertical scaling, as well as replica scaling. Meanwhile, the fast online DDL capability of the TDStore engine allows users to easily adjust database structures without affecting business operations as a response to various unpredictable business changes. This provides strong support for addressing modern complex business needs. The TDStore engine has already been adopted in the user activity data scenarios of Tencent's Mobile QQ Game Center.
Data Middle Platform
Although the traditional MySQL scheme of sharding databases and tables according to shard keys solves the core transaction problems in TP scenarios, it is not friendly to data operations or data aggregation scenarios (data serving/data hub). In these scenarios, queries are often conducted not by shard keys but by lightweight statistical analysis across multiple different dimensions. At this time, it is difficult to implement multi-dimensional queries simply, flexibly, and efficiently, whether through broadcast queries or index tables, while the TDStore engine is very suitable for such scenarios. The TDStore engine supports quickly aggregating data from multiple shards into a single instance, creating globally unique indexes, and conducting data consistency queries across multiple shards and dimensions. It meets the high-performance read-write requirements of data middle platform scenarios and supports users to perform data operations and analysis based on massive data. The TDStore engine has been adopted by Tencent Advertising for lightweight data warehouse scenarios.
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No
Feedback