Storage Services Overview:Tablestore

Scenarios

Tablestore can store petabytes of data in a single table. It also supports tens of millions of TPS and various index types, such as global secondary index, full-text index, inverted index, and spatio-temporal index. Tablestore is widely used for structured data in fields such as social networking, the Internet of Things (IoT), artificial intelligence (AI), metadata, and big data.

• Metadata

  When you store large amounts of data, such as documents and media files, it is important to also store and analyze the corresponding metadata. For example, in scenarios that involve storing e-commerce orders, bank account transaction histories, or phone bills from carriers, you also need to store and analyze a large amount of metadata. Tablestore helps you efficiently manage this metadata.

• Message data

  The proprietary Timeline model in Tablestore is designed for message data. It abstracts lightweight message queues that support many topics. You can use this model to store large amounts of social information, including instant messaging (IM) chats and feed stream information such as comments, posts, and likes. The Timeline model is used in many IM systems, such as DingTalk, to support the synchronization of large volumes of messages.

• Trajectory tracing

  Tablestore offers the Timestream model to manage and analyze trajectory data from scenarios such as running, riding, walking, and food delivery.

• Scientific big data

  Multi-dimensional gridded data is a type of scientific big data used in geoscience fields such as meteorology, oceanography, geology, and geomorphology. The volume of this data is large and growing. Scientists need to quickly browse data and run various types of online queries that require low latency. Tablestore can address the storage and query performance challenges associated with massive scientific big data.

• Internet big data

  Product designers for e-commerce and information platforms must aggregate, summarize, and analyze data from various platforms to guide future product development. Public relations and marketing departments also need to respond promptly to public sentiment. Tablestore helps you store and analyze tens of billions of records related to public sentiment on the Internet.

• Internet of Things

  Tablestore is ideal for storing time series data from IoT devices and monitoring systems. It supports direct SQL reads for big data analytics and provides efficient interfaces for incremental stream reads, enabling both batch analytics and real-time stream computing.

Performance

A single table in Tablestore can support up to 10 PB of data, trillions of records, tens of millions of TPS, and millisecond-level latency. It features automatic load balancing and hot spot migration, which eliminates the need for manual operations and maintenance (O&M). Tablestore delivers high write throughput and stable, predictable read and write performance. For more information, see the Tablestore performance white paper.

Data durability and service availability

Tablestore stores multiple backups of your data on different machines across different racks and quickly recovers data in the event of a failure. The service is designed for 99.99% high availability and 99.999999999% (eleven 9s) data reliability.

Scalability and elasticity

Tablestore uses data partitioning and load balancing to achieve seamless storage expansion. As the data volume in a table grows, Tablestore automatically adjusts the data partitions to allocate more storage to that table. Tablestore can scale to support at least 10 PB of data storage, with a single table supporting at least 1 PB of data or one trillion records.

Security

Tablestore provides table-level and API-level authentication and authorization mechanisms. It supports Security Token Service (STS) temporary authorization, custom permission authentication, and resource isolation for root accounts and RAM users. For more information, see Introduction to RAM and STS. Tablestore supports access from the Internet, ECS internal networks, and VPCs, and provides network access control.

Interfaces

Tablestore provides standard RESTful API operations. Developers typically use tools or SDKs that encapsulate these API operations to develop applications. Tablestore currently provides SDKs for various developer languages, including Java, Python, PHP, and Go. The command-line interface (CLI) provides simple commands for managing instances, tables, and data. It supports Windows, Linux, and Mac platforms.

In the Tablestore management console, you can create instances, data tables, and search indexes. You can also perform basic data read and write operations and monitor access metrics for instances and tables, such as queries per second (QPS), latency, and the number of requests.

Billing models

Tablestore offers a pay-as-you-go billing method that helps you handle traffic fluctuations and high-concurrency, low-latency requests at a low cost. You can also purchase subscription-based resource plans in advance, and usage fees are deducted from your plan's quota. Tablestore billing is based on usage. Billing items include data storage, read throughput, write throughput, and outbound Internet traffic. Additional fees are incurred if you use search indexes or global secondary indexes. For more information, see Billing overview.