Tablestore is a serverless table storage service for large volumes of structured data. It supports four data models: Wide Column, AI (Agent Memory), TimeSeries, and Timeline. Tablestore is suitable for scenarios that involve AI data, metadata, message data, and spatio-temporal data. It provides features such as search index, vector search, and SQL query. Tablestore seamlessly integrates with compute ecosystems such as MaxCompute, Flink, Spark, and Presto, and is compatible with mainstream AI frameworks such as Dify, LangChain, and LlamaIndex. Tablestore offers the MCP intelligent Agent architecture for AI chat applications and the IoTstore solution for IoT needs, enabling data storage and intelligent analysis for all scenarios.
Terms
Before using Tablestore, you need to understand the following basic concepts.
Term | Description |
Region | A region is a physical data center. Tablestore is deployed in multiple Alibaba Cloud regions. You can select a region to use the Tablestore service as needed. For more information, see Regions and zones. |
Read/write throughput | Read throughput and write throughput are measured in read capacity units and write capacity units. A capacity unit (CU) is the smallest billing unit for read and write operations. For more information, see Read/write throughput. |
Instance | An instance is an entity used to manage and use the Tablestore service. Each instance is equivalent to a database. Tablestore performs access control and resource metering for applications at the instance level. For more information, see Instances. |
Endpoint | Each instance has an endpoint. An application must specify the endpoint when it performs operations on tables and data. For more information, see Endpoints. |
Data lifecycle | Time to live (TTL) is a property of a data table. It specifies the retention period of data in seconds. Tablestore runs in the background to clear expired data. This reduces your data storage space and storage costs. For more information, see Data versions and time to live (TTL). |
Data storage models
Tablestore provides three data storage models: Wide Column, TimeSeries, and Timeline. You can select a model based on your scenario. For information about the features supported by different data storage models, see Features.
Model | Description |
Wide table model | This model is similar to the Bigtable and HBase models. It can be used in various scenarios, such as metadata and big data. It supports features such as data versions, TTL, auto-increment primary key columns, conditional updates, local transactions, atomic counters, and filters. For more information, see Wide Column model. |
TimeSeries model | This model is designed based on the characteristics of time series data. It can be used in scenarios such as IoT device monitoring, device data collection, and machine monitoring data. It supports features such as automatic creation of time series metadata indexes and a wide range of time series query capabilities. For more information, see TimeSeries model. |
Message model | This model is designed for message data scenarios. It can be used in message scenarios such as instant messaging (IM) and feed streams. It meets the requirements of message scenarios for message ordering, massive message storage, and real-time synchronization. It also supports full-text search and multi-dimensional composite queries. For more information, see Timeline model. |
Methods
You can use the Tablestore product in the following ways.
Method | Description |
Console | Alibaba Cloud provides a web service page that lets you easily operate Tablestore. For more information, see the Tablestore console. |
SDK | Supports mainstream development languages such as Java, Go, Python, Node.js, .Net, and PHP. For more information, see SDK Reference. |
Command line interface | Lets you operate Tablestore using simple commands. For more information, see Command line interface. |
Quick Start
You can use the console or command line interface to quickly perform operations on data tables in the Wide Column model or time series tables in the TimeSeries model. For more information, see Quick Start for the Wide Column model and Quick Start for the TimeSeries model.
Computing and Analytics
Tablestore supports computing and analysis through MaxCompute, Spark, Hive or HadoopMR, Function Compute, Flink, and Tablestore SQL query. You can select an analysis tool based on your scenario.
Analytical tool | Applicable model | Operation | Description |
MaxCompute | Wide table model | You can create a foreign table for a data table in Tablestore using the MaxCompute client, which lets you access the data in Tablestore. | |
Spark | Wide table model | When you use the Spark compute engine, you can access Tablestore through E-MapReduce SQL or DataFrame programming. | |
Hive or HadoopMR | Wide table model | Use Hive or Hadoop MapReduce to access data in Tablestore. | |
Function Compute | Wide table model | Use Function Compute to access Tablestore and perform real-time computing on Tablestore incremental data. | |
Flink |
| You can use Real-time Compute Flink to access source tables, dimension tables, or sink tables in Tablestore to perform real-time computing and analysis on big data. | |
PrestoDB | Wide table model | After connecting PrestoDB to Tablestore, use SQL to query and analyze data in Tablestore, write data to Tablestore, and import data into Tablestore through PrestoDB on Tablestore. | |
Tablestore search index | Wide table model | Search index uses inverted indexes and columnar storage to address complex multidimensional queries and statistical analysis for big data. When your business requires queries on non-primary key columns, multi-column combined queries, fuzzy queries, or analytics such as finding maximum/minimum values, counting rows, or grouping data, define these attributes as fields in a search index and use it to query and analyze your data. | |
Tablestore SQL query |
| SQL query provides a unified access interface for multiple data engines. With the SQL query feature, you can access Tablestore data to run complex queries and perform efficient analysis. |
Migration and synchronization
You can smoothly migrate and synchronize heterogeneous data to Tablestore. You can also synchronize Tablestore data to other services such as Object Storage Service (OSS).
Category | Data synchronization | Description |
Data import | Use the Tablestore Sink Connector to batch import data from Apache Kafka into a Tablestore data table or time series table. | |
Use Tunnel Service, DataWorks, or DataX to synchronize data from one Tablestore data table to another. | ||
Data export | Use DataWorks to export full data from Tablestore to MaxCompute. | |
Use DataWorks to export full or incremental data from Tablestore to OSS. | ||
Use the command line interface or DataX to download data directly to a local file. You can also use DataWorks to synchronize data to OSS and then download the data from OSS to a local file. |
More features
To control user access permissions, you can use Resource Access Management (RAM) to implement custom permissions. For more information, see Grant permissions to a RAM user using a RAM policy.
You can also further restrict user access permissions using control policies in a resource directory, Tablestore network ACLs, and Tablestore instance policies. For more information, see Authorization management.
To ensure data storage security and network access security, you can use methods such as data table encryption and VPC network access. For more information, see Data encryption and Network security management.
To prevent important data from being accidentally deleted, you can use the data backup feature to periodically back up important data. For more information, see Data backup.
To configure alert notifications for monitoring metrics, you can use Cloud Monitor. For more information, see Data monitoring and alerts.
To visualize data in forms such as charts, you can use DataV or Grafana. For more information, see Data visualization.