Quick start guide to Tablestore with real-world case studies

AI practices

A vector database is an important component of AI applications. Tablestore integrates with third-party open source frameworks such as LangChain, LlamaIndex, PAI-RAG, LangEngine, LangChain4J, Dify, and MCP. Compared with traditional databases and dedicated vector databases, Tablestore provides hybrid retrieval of vectors and scalars, supports petabyte-scale data processing with automatic horizontal scaling, and seamlessly integrates with the Alibaba Cloud big data ecosystem. These features allow you to easily and quickly build AI applications.

• Quick start for Tablestore MCP Service

• Cross-application AI memory: Tablestore OpenMemory MCP

• Quickly build a RAG application using Dify and Tablestore

• Deploy a Dify application integrated with Tablestore using Compute Nest

• Quickly build a RAG application using PAI-EAS and Tablestore

Metadata

Storing and analyzing file metadata is essential when working with large volumes of data, such as documents and media files. This also applies to the large amounts of metadata from e-commerce orders, bank transactions, and ISP bills. A single Tablestore table supports petabyte-scale storage, tens of millions of queries per second (QPS), and various index types, including global secondary index, full-text index, inverted index, and spatiotemporal index. These features meet various online query requirements across different scenarios and allow you to easily implement efficient data management.

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: Architecture

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: Data synchronization using DTS

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: Data synchronization using Canal

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: Order search

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: SQL query and analysis

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: Data processing using ETL

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: Historical data analysis

• Practices of a large-scale order system based on a MySQL and Tablestore tiered storage architecture: Data stream computing

• Build a billion-scale order management solution using Tablestore

• Build a billion-scale store search system using Tablestore search index

• Analysis of a mass intelligent metadata management system

• A mass insurance policy query platform based on Tablestore

Message data

The proprietary Timeline model of Tablestore is designed primarily for message data. It functions as a lightweight MSMQ that supports many topics and can store large amounts of social data, including IM chat messages and feed stream information such as comments, replies, and likes. The model provides simple and easy-to-use interfaces. The Tablestore Timeline model has been applied in numerous IM systems, for example, to support massive message synchronization in DingTalk. In addition, Tablestore uses the pay-as-you-go billing method, which cost-effectively meets requirements for significant fluctuations in access, high concurrency, and low latency.

• Message system architecture in modern IM systems: Architecture

• Message system architecture in modern IM systems: Model

• Message system architecture in modern IM systems: Implementation

• Tablestore Timeline: Easily build IM and feed stream systems with tens of millions of users

• Feed stream system design: An overview

• How to build a feed stream system with tens of millions of users

• How to optimize the architecture of a high-concurrency IM system

Trajectory tracing

Tablestore provides the Timestream model for trajectory scenarios. This model offers petabyte-scale storage, tens of millions of transactions per second (TPS), and millisecond-level latency. It also supports various index types, such as global secondary index, full-text index, inverted index, and spatiotemporal index. With the Tablestore time series model, you can easily manage and analyze trajectory data from activities such as running, cycling, walking, and food delivery.

• Tablestore time series data storage: Architecture

• Store massive sports trajectory data using Tablestore

• Drug regulatory system architecture revealed: Storage and query of mass traceability data

• Manage massive motorcycle trajectories using Tablestore

• Super Express: How to use a system to ensure on-time delivery

Scientific big data

Multidimensional grid data is a type of scientific big data widely used in earth science fields such as meteorology, oceanography, geology, and topography. As the volume of this data grows, scientists in these fields need to quickly browse data and perform various types of online queries with low latency. Tablestore is a serverless table storage service for massive structured data. It provides vast storage capacity, supports large-scale concurrent access, and delivers low-latency performance. Tablestore can easily resolve the storage and query performance issues associated with scientific big data.

A practical solution for massive meteorological grid data based on Tablestore

Internet big data

Trending news and entertainment topics can be forwarded tens of thousands of times and read by millions of people within minutes. The ability to understand public sentiment in real time and respond accordingly is crucial for many enterprises. In addition, the number of product orders on various e-commerce platforms and user purchase reviews also have a significant impact on other consumers. Product designers need to collect, summarize, and analyze data from various platforms to make decisions about future product development. Public relations and marketing departments also need to respond promptly to public opinion. A single Tablestore table provides petabyte-scale storage, tens of millions of QPS, and various index types. This helps you easily store and analyze tens of billions of records related to public opinion on the internet.

• A practical guide to unified stream and batch SQL processing with Tablestore and Spark

• Perform real-time statistics on transaction data using Tablestore and Flink

• Storage design for a public opinion analysis system with tens of billions of records

• Perform real-time aggregation on massive monitoring logs using EMR Spark Streaming SQL

IoT

A single Tablestore table provides petabyte-scale data storage without requiring sharding. It also supports tens of millions of QPS, which can easily meet the storage requirements for time series data from IoT devices and monitoring systems. The ability to directly read data for big data analytics using SQL, combined with efficient incremental stream reading interfaces, allows for easy batch analytics and real-time stream computing.

• A one-stop IoT storage solution based on Tablestore: Scenarios

• A one-stop IoT storage solution based on Tablestore: Table design

• A one-stop IoT storage solution based on Tablestore: Data operations

• A one-stop IoT storage solution based on Tablestore: Spark analytics

• A one-stop IoT storage solution based on Tablestore: Data lake analytics

• A one-stop IoT storage solution based on Tablestore: Flink real-time computing

• A car-sharing management platform based on Tablestore

• Architecture of a Wi-Fi device supervision system based on Tablestore

• IoT metadata management based on Tablestore

• How to efficiently store massive GPS data