Overview
Last updated: 2025-10-17 12:04:14
This document introduces a new feature of TDSQL-C for MySQL based on the LibraDB engine: the analysis engine.
Note：
TDSQL-C for MySQL supports two cluster types: transactional cluster and analysis cluster. The transactional cluster can provide HTAP capabilities by read-only analysis engine in conjunction with TXSQL read-write instances, while the analysis cluster only contains read-only analysis engine instances and synchronizes the configured data source data to the analysis engine via Zero-ETL capabilities.
Background
As a cloud-native database product, TDSQL-C for MySQL has undergone extensive optimization to support high concurrency, strong consistency, and enterprise-level database characteristics. It supports high-performance online transaction processing capabilities based on the TXSQL engine. However, in addition to using databases for high QPS online transactions, many business systems also require mining data and utilizing databases for data analysis, to help enterprises better make business decisions and drive the iterative innovation of businesses for quickly adapting to the market environment changes.
To support the high-performance online transaction processing capabilities and ensure the stability of business queries, traditional databases typically select row-based storage and adopt the Volcano model for execution, so they cannot efficiently serve analytical queries. Certainly, in some businesses, the traditional database + data warehouse solution is selected to support hybrid transaction/analytical processing, but this solution requires high maintenance costs. Moreover, customers need to build the ETL tools from the database to the data warehouse by themselves, and the real-time performance and consistency requirements for the data cannot be well satisfied.
Therefore, Tencent Cloud Native Database TDSQL-C for MySQL newly supports this analysis engine feature, providing you with efficient and real-time data analysis services.
What Is an Analysis Engine?
The analysis engine is a new feature supported by TDSQL-C for MySQL. This feature is implemented based on the LibraDB engine and provides services based on read-only instances. Its pluggable engine design allows for flexible creation and termination, and meanwhile provides you with massive data processing and efficient real-time complex analysis capabilities.
Supported Regions and AZs
Feature Strengths
High-Speed Analysis Engine LibraDB
The LibraDB engine supports complex query analysis on TB-level data with a very low execution latency, enabling your business analytics system to efficiently extract useful information from massive databases. The LibraDB engine supports vectorized engines, large-scale parallel execution, and other acceleration features for analytical queries. No matter in multi-table JOIN and data aggregation and sorting of super-large tables, or in complex nested SQL queries, the LibraDB engine can provide excellent performance experience.
Pluggable Analysis Engine
The LibraDB engine is compatible with MySQL protocols and syntax, allowing users to run complex queries directly in LibraDB without modifying their business logic. Users can enable the analysis engine based on actual business needs, and disable it at any time when analytical acceleration is not required, helping to control costs effectively.
Real-Time Columnar Data Loading Capability
With the built-in data synchronization component of the LibraDB engine, existing data in TDSQL-C for MySQL can be quickly loaded into the analysis engine. After the initial data load, all subsequent changes made to the data in the read-write instance can be synchronized in real time, ensuring consistency between row-based and columnar data. In addition, to address the inefficiency of data changes in traditional columnar storage under high-concurrency data update and deletion scenarios, the LibraDB engine offers columnar storage capabilities optimized for high-concurrency data updates, enabling real-time synchronization and achieving zero-latency performance.
Specified Data Loading Capability
In traditional read-only instances, all data from the primary database should be fully synchronized to the secondary database. However, with the analysis engine, specified objects can be loaded into the engine, rather than requiring all objects to be loaded. Users can choose to load only those databases and tables that need acceleration through the analysis engine, or those with analytical value, enabling flexible control over the disk space used by the analysis engine.
Multi-source data merge capability
In the product form of the analysis cluster, multiple data sources can be added to aggregate and synchronize data from multiple databases into the analysis engine. The capability to merge databases and tables helps provide efficient and real-time data analysis services.
Ultra-High Data Compression Rate
Based on the columnar storage structure, it provides both the ultra-high data scanning performance and an average compression ratio of 4-5 times, significantly reducing the storage costs.
Perfect Cloud Hosting Capabilities
Through fully hosted product design, you can experience the out-of-the-box data analysis capabilities, with no need to consider the complex ETL logic or backend database operations. Additionally, the comprehensive monitoring feature achieves meticulous filtering of core metrics from TXSQL to the analysis engine, and from the link layer to the storage layer, to simplify your operations, help you quickly understand the instance health status through key metrics, and provide effective optimization guide for the use of business systems. Furthermore, you can set custom threshold alarms to prevent potential exceptions in advance.
Applicable Scenarios
The analysis engine is designed to provide users with real-time, high-performance data analysis, helping to eliminate the complex Ops challenges of building custom ETL tools. With its built-in features, users can easily create data analysis instances with a single click, using them as a foundation for business decision-making and fully unlocking the value of their data.
Report Analysis and Real-Time Dashboard
For report systems designed for internal enterprise analysis and management, users can view the real-time operational status of online business systems. It also applies to data analysis tasks for business operations. In such scenarios, the SQL queries are complex and variable in pattern, requiring high throughput and involving large volumes of online data. The analysis engine meets the real-time and high-performance requirements of these types of workloads.
User Reputation and Behavior Analysis
In advertising and game operations scenarios, in-depth analysis of user behavior and user profiling is often required, with the results used to support real-time business decisions. These scenarios typically involve large volumes of data, require timely responses, and have high query QPS. By using the analysis engine, users can quickly obtain the necessary data for analysis, enabling accurate insights into user behavior that serve as a decision-making foundation for precise business targeting.
Real-Time Data Warehouse
The analysis engine can also be used in scenarios such as order analysis during major e-commerce promotions, waybill analysis in the logistics industry, performance analysis and metric computation in the financial sector, live streaming quality analysis, ad delivery analysis, intelligent dashboards, and probe analysis, providing ultra-high performance for complex queries.
Big Data Reconciliation and Batch Computing
In certain online services, especially those involving financial transactions, periodic data aggregation and reconciliation are required. Performing batch reconciliation on traditional row-based data is often inefficient and resource-intensive, making it difficult to meet business expectations in a timely manner. By leveraging the high-concurrency computing capabilities of the analysis engine, users can fulfill these business needs with extremely high efficiency.
Aggregation analysis
In certain online businesses, data from multiple modules should be aggregated and queried. In sharded database and table scenarios, decentralized data should be merged. With the capabilities of an analysis cluster, data can be easily merged and aggregated for configuration. Users do not need to self-build complex synchronization linkages, and can fulfill business needs with high efficiency.
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No
Feedback