:Architectures

System architecture

The following figure shows the system architecture of Tablestore.

Scenarios

Tablestore is suitable for building systems that handle metadata, message data, spatio-temporal data, and big data.

Data access

You can access Tablestore by using an SDK, DataWorks, or an IoT rules engine. Tablestore can be used to store a variety of structured data, including application data, message data, and IoT data.

Tablestore

• Multiple data storage models

  Tablestore provides three data storage models: the Wide Column model, the TimeSeries model, and the Timeline model.

  Model	Description
  Wide Column	This model is similar to the Google Cloud Bigtable and HBase models, and can be used in multiple scenarios such as metadata and big data storage. The Wide Column model supports features such as max versions, TTL, auto-increment primary key column, conditional update, local transaction, atomic counter, and filter. For more information, see Wide Column model.
  TimeSeries	This model is designed to store time series data generated from multiple scenarios such as IoT device monitoring, device data collection, and machine monitoring. The TimeSeries model supports automatic indexing of time series metadata and time series query by various composite conditions. For more information, see TimeSeries model.
  Timeline	This model is designed to store message data and is suitable for storing message data generated from IM applications and feed streams. This model can meet the specific requirements of messaging processes, such as message order preservation, storage of large numbers of messages, and real-time synchronization. This model also supports full-text search and Boolean query. For more information, see Timeline model.

• Powerful data indexing capabilities

  Tablestore provides powerful data query features, such as primary key-based query, secondary index-based query, and search index-based query.

  Index type	Description
  Primary key of data tables	A data table is similar to a large map. Data tables support queries based only on primary keys.
  Secondary index	You can create one or more index tables and perform queries by using the primary key columns of the index tables.
  Search index	The search index feature uses inverted indexes, Bkd-trees, and column stores to cater for a variety of queries, such as non-primary key column-based query, Boolean query, geo query, full-text search, fuzzy query, nested structure query, and statistical aggregation.

• Hot and cold data tiering

  Cold data and hot data are automatically tiered. Tablestore provides two types of instances, namely, high-performance instances and capacity instances, to meet the data storage requirements of various business scenarios.

  Instance type	Description
  High-performance instance	This instance type is suitable for scenarios that require high read and write performance and high concurrency, such as gaming, financial risk control, social networking applications, and recommendation systems.
  Capacity instance	This instance type is suitable for business that is cost-sensitive but does not have high requirements on read performance, such as business that involves log monitoring data, Internet of Vehicles (IoV) data, device data, time series data, logistics data, and public opinion monitoring data.

• Data delivery

  Data is fully backed up or delivered in real time to Object Storage Service (OSS). Data delivery is compatible with open source ecology standards and the naming conventions followed by Hive. Delivered data is stored in the Parquet format. You can use the Data Lake Analytics (DLA) and E-MapReduce (EMR) services to analyze tables that are delivered to OSS.

Compatibility with the computing ecosystem

• Tablestore is compatible with mainstream open source stream and batch computing engines, such as Flink, Spark, and Presto.

• Tablestore seamlessly integrates with Alibaba Cloud big data platforms, including DataWorks, DataHub, and MaxCompute.

Typical application architectures

Tablestore provides services based on three typical application architectures: the Internet application architecture, data lake architecture, and IoT architecture.