All Products
Search

This Product

    Object Storage Service:Tablestore

    Document Center

    Object Storage Service:Tablestore

    Last Updated:Mar 20, 2026

    Tablestore is a data storage service developed by Alibaba Cloud to store a large amount of structured data. It stores petabytes of data with tens of millions of transactions per second (TPS) and millisecond latency. Compatible with HBase, it provides three data models—Wide Column, Timeline, and Timestream—along with global secondary index, full-text search, inverted index, and spatio-temporal index for flexible query patterns.

    Use cases

    Tablestore stores petabytes of data in a single table and handles tens of millions of TPS across four index types, making it suitable for a wide range of structured data workloads.

    • Metadata management E-commerce orders, bank transaction histories, and phone billing records generate large volumes of metadata that must be stored, queried, and analyzed at high throughput. Tablestore's wide-column model and secondary index support efficient metadata storage and lookup at this scale.

    • Message and social data The Timeline model provides lightweight, high-throughput message queues that scale to a large number of topics. It powers instant messaging (IM) applications—including DingTalk—as well as social feeds (comments, posts, and likes), enabling real-time synchronization of large message volumes.

    • Trajectory tracing The Timestream model is purpose-built for trajectory data. It handles diverse movement scenarios—running, riding, walking, and food delivery—and supports both management and analysis of trajectory datasets.

    • Scientific big data Gridded data used in geoscience fields (meteorology, oceanography, geology, and geomorphology) grows rapidly and requires fast browsing with low-latency online queries. Tablestore meets the storage capacity and query performance requirements of these scientific workloads.

    • Internet big data E-commerce and content platforms need to collect and analyze behavioral data across product lines, while public relations teams monitor and respond to public opinions in near real time. Tablestore handles tens of billions of public opinions to support these analysis pipelines.

    • IoT time series data Tablestore stores time series data from Internet of Things (IoT) devices and monitoring systems. Direct SQL access and incremental data stream APIs let you run offline batch analysis and real-time stream computing on the same dataset.

    Performance

    Tablestore stores tens of petabytes of data and trillions of records in a single table, with tens of millions of TPS and millisecond latency. It handles automatic load balancing and hotspot migration without manual operations and maintenance (O&M), delivering high write throughput and predictable read and write performance. For benchmark details, see the Tablestore performance white paper.

    Data durability and service availability

    Tablestore creates multiple backups of data and stores them in different servers across racks. When a backup fails, Tablestore immediately uses another backup to restore the data. This mechanism guarantees data durability of 99.99% and service availability of 99.999999999% (eleven 9s).

    Scalability

    Tablestore uses shards and load balancing for seamless horizontal scaling. As data in a table grows, Tablestore automatically adjusts partition sizes—no manual intervention required. A single Tablestore deployment holds a minimum of 10 PB of data; a single table holds a minimum of 1 PB or one trillion records.

    Security

    Tablestore enforces authentication and authorization at the table and operation level. It supports Security Token Service (STS) temporary authorization, custom authentication, and Resource Access Management (RAM) users for resource isolation. Network access is available over the Internet, from ECS instances, and through VPCs, with network access control across all access paths. For details, see RAM and STS.

    Access methods

    Tablestore exposes a standard RESTful API. Build applications using the Tablestore SDKs for Java, Python, PHP, and Go, or use the TableStore CLI—a command-line interface (CLI) tool for Windows, Linux, and Mac that covers table operations, single-row operations, simple stress testing operations, and data backup.

    The Tablestore console lets you create instances, tables, and search indexes, run basic read and write operations, and monitor queries per second (QPS), latency, and request counts for instances and tables.

    Billing

    Tablestore supports two billing methods:

    • Pay-as-you-go: pay only for the resources you use. Handles traffic spikes and high-concurrency, low-latency workloads without upfront commitment.

    • Subscription: purchase resource plans in advance to cover predictable usage at a lower effective rate.

    Billable items include storage usage, read throughput, write throughput, and Internet outbound traffic. Using the search index or global secondary index feature incurs additional fees. For a full breakdown, see Billing overview.

    What's next