Feature Introduction
The columnar read-only analysis instance is primarily designed for efficient analytical queries. It serves as an extended read-only analysis component that provides customers with real-time, high-performance processing of complex SQL. By leveraging its columnar storage capabilities, vectorized parallel execution engine, and optimizer enhanced for distributed parallel execution, the instance enables customers to easily experience efficient analytical capabilities directly within the database. Furthermore, its columnar storage is specifically optimized for high-QPS updates and transaction ACID properties, ensuring real-time and consistent query data.
Principles
The kernel of the columnar read-only analysis instance automatically converts data from the primary node into columnar storage and synchronizes binlog in real time, ensuring that the data remains consistent with the primary node. It then processes complex user SQL and accelerates query execution by leveraging parallel computing capabilities and a vectorized execution engine.
Supported Features
The kernel of the columnar read-only analysis instance supports a variety of excellent features. The following section briefly introduces the features supported by the product.
Parallel Computing Capability
Within the columnar read-only analysis instance, the computational resources of the nodes can be fully utilized to execute user SQL. When SQL runs in the columnar read-only analysis instance, a single SQL query task can be split into multiple subtasks, and these subtasks are allowed to run concurrently across multiple processors. This multi-threaded execution mode effectively improves CPU utilization and IO resource efficiency, thereby achieving the goal of accelerated processing.
Columnar Storage for High-Speed Change Scenarios
The columnar read-only analysis instance, through optimizations and support at the storage layer, can meet the data consistency requirements for user data changes in high-concurrency scenarios. This prevents missing the analysis window due to data latency caused by frequent data changes on the read/write instance.
Specifying Data Loading Capability
Not all data in the TDSQL Boundless primary instance has analytical value, so it is unnecessary to load all objects into columnar storage. Therefore, the columnar read-only analysis instance supports the capability of loading specified objects. You can specify the objects to be loaded for the columnar read-only analysis instance either in the data loading console or via command-line SQL.