Tair  Play Video

Tair is a Redis-compatible in-memory database service that provides a variety of data structures and enterprise-level capabilities such as Global Distributed Cache, data flashback, and Transparent Data Encryption (TDE).


Tair is an in-memory database service based on ApsaraDB for Redis Community Edition and brings enterprise-level features and multiple storage media with cost-effectiveness.

DRAM-based instance: delivers ultra-high throughput and triples the performance of open source Redis, which is suitable for industries such as livestreaming, e-commerce, and online education.
Persistent memory-optimized instance: implements real-time data persistence at 70% the cost of ApsaraDB for Redis Community Edition, which is an ideal choice for the gaming industry.
ESSD-based instance: provides storage capacities of up to hundreds of terabytes, which is suitable for the storage of large amounts of warm and cold data with high performance.

Free Trial

See Special Offer Terms & Conditions for specifics and limitations.

Tair (Enhanced Redis)

Fully compatible with Redis, the most popular KV caching database.


Redis 5.0/6.0

DRAM-base Edition

1GB Performance-enhanced

New User Only
30 Days Free


Full Compatibility with Redis and Instance Upgrade Without Downtime

Tair is fully compatible with Redis. Migration to and from both databases can be performed out-of-the-box without any code modifications. Additionally, Tair is bundled with a variety of extended data structures that can simplify data storage and development complexity. The configuration change feature of ApsaraDB for Redis Community Edition can be used to upgrade an ApsaraDB for Redis instance to a Tair instance without downtime. View Details

Data Persistence

Tair persistent memory-optimized instances use persistent memory and support command-level real-time persistence, semi-synchronization, and data restoration to a specific point in time. This instance series provides 90% read and write capabilities of ApsaraDB for Redis Community Edition at only 70% of the cost. This makes the instance series an ideal choice to replace solutions that use open source Redis for caching and ApsaraDB RDS for persistent storage. This way, the simplified database system ensures data consistency between caching and persistent data storage. View Details

Data Restoration to a Specific Point in Time

Tair provides the data flashback feature that can restore data to a specific point in time, accurate to the second. This feature enables you to restore an entire instance or specific keys to a new instance or the current instance of a specific point in time. This fine-grained data restoration capability prevents data loss caused by accidental operations and ensures data security. View Details

Global Distributed Cache and Data Synchronization

Global Distributed Cache is an in-house active geo-redundancy database system. The cloud-native geo-replication capability of Tair supports data synchronization across up to three regions. This helps you implement geo-redundancy, disaster recovery, data consistency across regions, and nearby access with reduced latency. View Details

Multiple Extended Data Structures

Tair provides a wide range of built-in extended data structures on top of the support for all open source Redis data structures. These extended data structures accelerate development timelines and improve code security by sparing you large amounts of code and the use of Lua scripts. Additionally, these data structures enable you to implement high-performance distributed locks, multidimensional leaderboards, sliding windows, and JSON file storage with ease. View Details


Ease of Use: A Broad Set of Data Structures to Set up Real-time Scenarios

Provides multiple instance series tailored towards hot, warm, and cold data storage, high performance, low costs, and data persistence.

Tair DRAM-based Instances

Uses a multi-threading model to increase the queries per second (QPS) of each node and handle a large number of connections with three times performance of open source Redis for requirements on high concurrency and low latency. View Details

Tair Persistent Memory-optimized Instances

Provides native command-level real-time data persistence and high performance for storage of large amounts of data with cost-effectiveness (90% throughput and 70% costs compared with ApsaraDB for Redis Community Edition). View Details

Tair ESSD-based Instances

Leverages Alibaba Cloud ESSDs to cater for massive data storage and query with up to 85% cost reduction compared with ApsaraDB for Redis Community Edition, and provides native Redis command-level data persistence and compatibility with native Redis data structures and APIs. View Details

Multiple Architectures: Standalone, Master-replica, Read/Write Splitting, and Cluster

Standalone for caching-only scenarios, master-replica for automatic failover, read/write splitting for read-intensive scenarios, and cluster for capacity-sensitive scenarios

Standalone Architecture

Designed for caching-only scenarios and scales out smoothly to deliver cost-effective, high QPS performance.

Master-replica Architecture

Provides a master node and a replica node for real-time data synchronization. Automatic failover completes within seconds, and is completely transparent, ensuring the availability of your applications.

Cluster Architecture

Adopts a distributed architecture in which each shard contains a master node and a replica node to implement automatic failover and ensure high availability, provides multiple specifications, and supports push-button scaling.

Read/Write Splitting Architecture

Designed for read-dominant scenarios with a maximum of 100,000 writes per second and 600,000 reads per second to break bottlenecks of hotkey reads.

Enterprise-level Capabilities: Global Distributed Cache, Data Flashback, and Multiple Data Structures

Enterprise-level features for fast implementation in various scenarios

Data Restoration to a Specific Point in Time

Allows you to restore data to a specific point in time within the last seven days by second to prevent data loss caused by accidental operations and improve data security. View Details

Real-time Data Persistence

Implements data persistence without dependence on append-only files (AOF) and Redis Database (RDB) files to prevent data loss from power outages and enhance data reliability. View Details

Global Distributed Cache

Supports real-time read and write synchronization across zones and regions, provides cross-region replication, and achieves geo-redundancy, disaster recovery, and nearby access with reduced latency. View Details

Automatic Scale-up

Handles traffic spikes and prevents out-of-memory errors to ensure business stability.

Accelerated Hotkey Read

Uses proxy nodes to cache hotkeys and resolve hotkey-related performance bottlenecks. You no longer have to manually modify your code or predict resource usage. View Details

Enhanced Data Structures

Provides Redis-compatible enhanced data structures to implement multidimensional leaderboards, distributed locks, Bloom filters, and trajectory monitoring for agile development. View Details

Elasticity: Multidimensional Scalability and Transparent Upgrades

One-click configuration change to improve instance capacities and performance or change instance architectures

Scalable Storage

Provides a wide range of memory configurations, allowing you to scale as your business changes.

Scalable Performance

Supports memory and throughput scaling to achieve tens of millions of queries per second.

Flexible Configurations

Flexibly switches between standalone and master-replica architectures for in-memory storage or persistent storage capabilities.

Intelligent O&M: Purpose-built Monitoring and Management Platforms Provide Push-button Troubleshooting

Provides user-friendly graphical database management and O&M platforms. This helps reduce O&M costs and support real-time and historical analysis of large keys and hotkeys.

Monitoring Platform

Monitors instance metrics such as CPU utilization, memory usage, number of connections, and disk usage in real time and raises alerts to track your instance status.

Visualized Management Platform

Clones, backs up, and restores instances with one click.

Visualized DMS Platform

Provides the Data Management (DMS) console for visualized management and enhanced development and maintenance efficiency.

Powerful Monitoring Capabilities

Provides monitoring within seconds for storage and latency features such as real-time hotkeys and large keys, latency insights, and audit logs.

Comprehensive Diagnostic Reports

Provides checks for request latency and instance health with automatic or manually generated diagnostic reports to identify issues.


High Scalability

Livestreaming applications constantly experience traffic spikes. Tair cluster instances can be used to meet the needs for high concurrency, low latency, and up to millions of QPS.


  • Transparent Elastic Scaling

    Handles system access, user visits, and traffic spikes without affecting your business.

  • Data Security and Stability

    Each shard in a cluster instance uses a master-replica architecture to implement disaster recovery and ensure data security.

Meets the needs for duplicate servers, cross-server data transfer, and elastic scaling.

Gaming applications have high requirements for simplified storage solutions, scalability, and high performance. Tair effectively covers these needs so you can focus on developing your applications.


  • Data Persistence

    Gaming applications have requirements for duplicate servers, cross-server data transfer, and elastic scaling. In this case, data persistence can decrease the load on the database engine and simplify development.

  • One-click Cloning

    Meets the needs for duplicate servers of gaming applications and allows you to clone instances by using backup files.

  • Disaster Recovery

    Provides free seven-day retention period and one-click restoration for backup files to ensure data security.

Seamless Integration with Big Data Computing

E-commerce promotions such as flash sales exert heavy load on the inventory management system. In this case, Tair can handle read requests with ease and provides data persistence.


  • High QPS

    Uses a multi-threading model to handle up to 300,000 QPS and high concurrency scenarios.

  • Data Persistence

    Provides point-in-time snapshots for AOFs and RDB files to ensure data persistence.

  • Optimized Storage Architecture

    Uses a storage architecture that combines Tair with ApsaraDB RDS and provides shared caching and data persistence storage to improve database performance.

Caching Scenarios

Web applications are read-and-write-intensive and cost-sensitive. Tair standalone instances meet these requirements with caching capabilities, high QPS, and comprehensive data persistence.


  • High Read and Write Performance

    Tair cluster instances that use the multi-threading model can handle up to 300,000 QPS and massive traffic.

  • High Availability

    Protects your business from single-node failure.

  • Cost-effectiveness

    Tair standalone instances deliver cost-effective, caching-only performance.

Special Offer Terms & Conditions

The offers detailed on this page are of a limited time and are valid for as long as the page shows. The offers apply only to purchases of the specific products listed on this page from Alibaba Cloud. They do not apply to purchases made at third-party resellers.


These are the terms and conditions that apply to promotions on alibabacloud.com, including ones where you must select products as part of the offer. No alternative terms and conditions apply.
• Alibaba Cloud reserves the right to modify or cancel the offer at any time.
• These offers only apply to products and digital content.
• These offers are limited to one per customer and account. If you have multiple accounts, only one account is eligible to participate.
• Only users that have passed our Real Name Registration verification process are eligible to participate.
• First instance promotion is applicable only to users who have never bought an instance under the same product. Please note that in this offer, PolarDB for MySQL, PostgreSQL, and Oracle are all categorized under the same product. Therefore, a user can only enjoy the promotion for only one of the following PolarDB databases: MySQL, PostgreSQL, or Oracle, and only if the user has not bought any instances under the PolarDB family.
• A user is eligible no more to purchase subscription instances with the promotion after buying first instance under the same product.
• Each of these offers may not be combined with other offers.
• Each of these offers are non-transferrable and may not be resold.
• Discounts and promotions will apply automatically when you check out if all terms and conditions have been met.
• If you violate any of the offer terms, the offer will be made invalid.

Show more

phone Contact Us