Tablestore is a cloud-native, serverless, and distributed NoSQL database that provides unified multi-model data storage. Tablestore integrates with the Alibaba Cloud and open source AI ecosystems to manage data storage and processing for complex business scenarios.
Unified multi-model storage
Tablestore supports multiple data models to meet the data storage requirements of various business scenarios, providing a unified, multi-model solution.
Wide table model: Based on the Bigtable/HBase architecture, this model supports a schema-free data structure. It is ideal for storing large amounts of structured and semi-structured data, such as metadata, orders, and messages.
Time series model: Optimized for scenarios such as the Internet of Things (IoT) and monitoring. This model provides efficient storage, compression, and query capabilities for time series data.
Message model: Designed for social networking scenarios, such as instant messaging (IM) and feed streams. This model supports message reading, writing, synchronization, and retrieval.
Indexing and query capabilities
In addition to primary key queries, Tablestore provides various indexing capabilities to support complex data retrieval.
Search index: Supports inverted indexes and provides query features similar to Elasticsearch. These features include full-text search, multi-field composite queries, range queries, and geo queries.
Core highlight - Vector search: Tablestore includes a built-in, professional-grade vector search engine. It supports hybrid searches that combine vector, full-text, and scalar searches. This is ideal for scenarios such as retrieval-augmented generation (RAG) and Agent Memory. It can store and retrieve tens of billions of vector data entries and seamlessly integrates with the AI ecosystem.
Secondary index: Supports global and local secondary indexes, which enable fast point queries and range queries based on attribute columns.
SQL query: Compatible with SQL query syntax to simplify development. You can use SQL to query and analyze data.
Deep ecosystem integration
Tablestore integrates with the Alibaba Cloud data ecosystem and mainstream open source ecosystems to enable seamless data forwarding.
AI ecosystem: Compatible with mainstream AI developer frameworks, such as Dify, LangChain, and LlamaIndex. Tablestore can serve as a memory storage backend to support building AI applications.
Compute and analytics overview:
MaxCompute/Hive/Spark: Read and write Tablestore data directly through foreign tables or connectors for offline batch processing.
Flink/Blink: Use Tablestore as a source and sink for Flink to perform real-time stream computing.
Function Compute (FC): Use Tablestore triggers for event-driven, serverless computing.
Presto: Use Presto for interactive data analytics.
Data tunnels and forwarding
Tunnel Service: Provides real-time subscription to and consumption of full and incremental data, which supports the construction of real-time data warehouses.
Data lake delivery: Automatically deliver Tablestore data to an Object Storage Service (OSS) data lake. The data is automatically converted to formats such as Parquet. This process simplifies integration with data lake analytics engines, such as Hive, and enables hot-cold data separation for low-cost archiving.