Tair is a cloud-native in-memory database service that is developed in-house by Alibaba Cloud. It is compatible with the open source Redis protocol and provides Tair extended data structures and enterprise-class features. This topic describes the functions and features of Tair.
Engine versions and instance types
Tair is compatible with multiple Redis versions and provides multiple instance architectures that are suitable for various business requirements.
Versions
Tair is compatible with Redis 4.0, 5.0, and 6.0.
Architectures
Tair provides the standalone, master-replica, read/write splitting, and cluster architectures.
Instance specifications
Instances that do not have the cluster mode enabled provide a memory capacity of 256 MB to 64 GB. Instances that have the cluster mode enabled provide a memory capacity of 2 GB to 16 TB.
Storage types
Tair provides instances of various storage types, performance levels, and costs to meet the requirements in different scenarios.
DRAM-based instance
Provides twice to three times the performance of open source Redis and four times the capability of open source Redis to execute Lua scripts.
Persistent memory-optimized instance
Provides the same performance as open source Redis at only 70% cost, and delivers command-level data persistence.
ESSD-based instance
Uses enhanced SSDs (ESSDs) as the storage media and provides 30% to 50% of the performance of open source Redis at only about 15% cost. ESSD-based instances are suitable for storing large amounts of warm and cold data. The cost-effectiveness is proportional to the data volume.
Enterprise-class capabilities
Tair provides various enterprise-class capabilities that open source Redis does not have to help your business navigate exceptional circumstances and improve development efficiency.
Extended data structures
Make your business more performant at significantly reduced development costs.
Proxy query cache
Automatically analyzes and caches the requests and responses for hotkeys. If the same requests are received from a client, the proxy nodes retrieve the responses to the requests from the cache and send the responses to the client. This helps accelerate responses to requests for hotkeys.
Global Distributed Cache
Provides an active geo-redundancy architecture that suits business scenarios in which multiple sites in different regions provide services at the same time. You can use this feature to deploy child instances across three regions. All child instances are readable and writable.
Data flashback based on point-in-time recovery (PITR)
Logs write operations to implement incremental backup that can be used to restore instance data to a specific point in time.
Instance management
Tair allows you to manage instances throughout their lifecycle, including creation, configuration change, destruction, and configuration of various projects.
Instance creation
Allows you to create Tair by using multiple methods, such as the Tair console, API operations, and Resource Orchestration Service (ROS).
Instance name customization
Allows you to specify custom instance names that are easy to identify.
Connection management
Supports private and public endpoints for a Tair instance and changes to instance endpoints and port numbers.
Tag management
Allows you to configure and modify tags for instances.
Instance configuration change
Allows you to change the configurations of a Tair instance without downtime, such as the instance architecture, specifications, number of read replicas, and number of shards.
Instance minor version update
Supports minor version updates without downtime for instances and proxy nodes to deliver more features and better stability and performance.
Instance migration across zones
Supports instance migration across zones within the same region. Tair also supports the deployment mode change between single-zone deployment and dual-zone deployment.
Manual master-replica switchover
Supports manual master-replica switchover to respond to business exceptions and planned fault drills.
Proxy node restart
Supports manual restart for proxy nodes to respond to fault drills and exceptions on proxy nodes.
Event center
Displays all events within your instance lifecycle such as configuration changes and instance migrations across zones.
VPC change
Allows you to change the virtual private cloud (VPC) of an instance within the same zone.
Configuration and management of instance parameters
Allows you to modify the values of instance parameters to respond to changing business requirements.
Instance restart
Supports manual restart for instances to respond to unexpected performance problems or connection issues caused by business exceptions.
Maintenance window
Allows you to set a maintenance window for an instance so that the system can maintain the instance and update the instance minor version. We recommend that you configure a time range during off-peak hours for the maintenance window to minimize the effects on your business.
Manual release for pay-as-you-go instances
Allows you to release a pay-as-you-go instance.
Recycle bin
Provides the recycle bin for instances that are released or have expired. You can re-create or destroy an instance in the recycle bin.
Security policies
Tair provides multiple access management policies, data encryption, and end-to-end encryption for connections to ensure data security across the board.
Logon control
Provides multiple logon methods for instances, including username and password logon, password-only logon, and password-free access over a VPC.
IP address whitelist
Allows you to configure IP address whitelists for an instance to manage the clients that can connect to the instance.
Account and permission control
Allows you to create accounts for an instance. You can separately configure the username and password for an account and manage the permissions of the account. You can grant read-only or read/write permissions to an account.
Transparent Data Encryption (TDE)
Uses the bring your own key (BYOK) method to encrypt data in the memory when the data is written to disks to ensure data security.
Disable high-risk commands
Allows you to disable high-risk commands based on your business requirements to enhance instance security.
Release protection
Prevents pay-as-you-go instances that have the release protection feature enabled from being released.
Data management in a visualized manner
Data query and write
Allows you to query and write data in a visualized manner, including inserting, deleting, and modifying data.
Monitoring and alerting
Tair provides multiple monitoring metrics and is integrated into CloudMonitor to implement alerting capabilities. This enables you to analyze instance performance and status across various dimensions.
Performance monitoring
Provides various metrics for you to monitor instance performance across the board.
Custom alert policy and notification method
Provides custom alert policies for a single metric or multiple metrics under multiple conditions. You can also configure different alert notification methods such as messages, emails, internal messages, and DingTalk.
Performance trends
Allows you to query the long-term performance trends of an instance.
Real-time performance
Allows you to query the real-time performance of instances.
Offline key analysis
Provides offline key analysis to identify large keys without additional workloads for your instance.
Real-time key statistics
Allows you to track requests in real time to analyze large keys and hot keys. This helps you quickly identify issues.
Slow queries
Allows you to log and monitor slow queries for instances.
Latency insights
Allows you to collect the latency statistics of all commands and custom events accurate to microseconds. This helps you troubleshoot faults and performance degradation for your Tair instance.
Log management
Tair provides multiple log types to help you analyze instance status.
Slow query log
Logs slow queries that last for a time period longer than the slow query log threshold. You can specify the slow query log threshold.
Active log
Provides active logs for instances.
Audit log
Works with Simple Log Service to provide audit logs for Tair instances. You can query, analyze, and export audit logs to track instance performance and status.
Disaster recovery and backup
Tair provides multiple disaster recovery methods to cope with instance failures, including single-zone deployment, dual-zone deployment, and Global Distributed Cache. Tair provides full backup and restoration and incremental backup and restoration. Tair supports manual and automatic backup policies.
Single-zone deployment
Deploys the master and replica nodes of an instance in the same zone for disaster recovery.
Dual-zone deployment
Deploys the master and replica nodes of an instance in different zones for disaster recovery across zones.
Global Distributed Cache
Provides an active geo-redundancy architecture that suits business scenarios in which multiple sites in different regions provide services at the same time. You can use this feature to deploy child instances across three regions. All child instances are readable and writable.
Master-replica switchover
Switches workloads from the master node to the replica node of an instance to cope with instance failures. The failover typically requires 30 seconds.
Full backup and restoration
Creates snapshots for instance data to implement full backup and restoration.
Incremental backup and restoration
Logs write operations to implement incremental backup that can be used to restore instance data to a specific point in time.
Backup policies
Supports manual and automatic backup policies. Automatic backup is executed once a day and cannot be disabled.