Storage
Redundant storage
Alibaba Cloud Object Storage Service (OSS) offers both locally redundant and zone-redundant storage to ensure data durability and availability across single or multiple zones. Locally redundant storage replicates data within a single zone across multiple devices and facilities, safeguarding against hardware failures. Zone-redundant storage extends this protection across multiple zones within the same region, maintaining data accessibility even if one zone becomes unavailable.
Cross-region replication
The cross-region replication feature provided by OSS automatically and asynchronously replicates files (objects) across different OSS data centers (regions). This replication includes operations such as creation, update, and deletion of objects, facilitating cross-region disaster recovery. For users who require high data security and availability requirements, this feature ensures that backup data from another OSS data center can be activated following a major disaster. Elastic Block Storage disks also support asynchronous replication of block storage data across regions or zones, providing disaster recovery backup for critical business data.
Versioning
Versioning is a bucket-level feature that protects data by storing previous versions of objects when they are overwritten or deleted. This lets users restore objects to any previous version in case of accidental overwrites or deletions.
Scheduled backup
You can schedule periodic backups of OSS data to Cloud Backup with the scheduled backup feature, then restore an object if you lose it. File Storage NAS integrates seamlessly with Cloud Backup. When backing up General-purpose NAS, instead of creating file system snapshots, Cloud Backup employs an efficient file system scan mechanism. Backup policies can be configured to create multiple data replicas, which can be used to promptly restore files in case of data loss or damage.
Consistency verification for replicated data
Elastic Block Storage provides the replication pair-consistent group feature that allows unified management and operations for asynchronous replication in disaster recovery scenarios. This ensures that data on multiple disks within the same group can be restored to a consistent point in time, providing instance-level or multi-instance-level disaster recovery protection.
Databases
Data disaster recovery
Alibaba Cloud databases offer robust backup and recovery capabilities, with automatic backup enabled by default for most common database products. Users can restore database instances using backup sets within the same region and zone or perform geo-disaster recovery to other regions and zones. Alibaba Cloud also provides standalone data disaster recovery products for fine-grained backup and recovery policy management.
Disk redundancy
Alibaba Cloud disk version databases leverage disk storage capabilities to ensure data reliability through multiple replicate redundancy. High-availability versions also provide redundancy for primary and secondary nodes.
Zone-disaster recovery
Alibaba Cloud database products, excluding single-node basic edition instances suitable for testing, offer high-availability features for primary and secondary nodes. Real-time data synchronization between these nodes, coupled with near real-time anomaly detection and high-availability switchovers, ensures zone-disaster recovery capabilities. Multi-zone deployment options further enhance this protection.
Geo-disaster recovery and global active-active
Geo-disaster recovery is primarily based on real-time data transmission, supported by Alibaba Cloud's low-latency internal network. Data Transmission Service (DTS) products enable real-time synchronization of mainstream relational databases, such as Alibaba Cloud RDS and PolarDB, across regions. Some products support two-way data synchronization and multi-write designs for cost-effective disaster recovery. DTS also supports products such as Redis, MongoDB, and PolarDB-X, while data migration products support DB2, Teradata, and HBase. PolarDB's Global Database Network (GDN) natively provides geo-disaster recovery capabilities through data synchronization across multiple clusters in different regions of a country.
PolarDB's Global Database Network (GDN) natively implements data synchronization for all clusters through multiple clusters across regions, providing geo-disaster recovery capabilities through GDN.
Flashback query
PolarDB's flashback query capability lets users quickly locate and recover from anomalies, such as accidental data deletion. This feature enables point-in-time flashback queries to swiftly restore accidentally deleted data, greatly improving fault recovery efficiency.