Data Management (DMS) is built on a three-layer service architecture: data sources, data operation and management, and commercial application. DMS also includes a built-in Data Disaster Recovery module that provides full backup, incremental backup, geo-redundancy, and data restoration.
Data sources
DMS connects to the following database types:
Relational databases: MySQL, SQL Server, PostgreSQL, PolarDB for MySQL, PolarDB for PostgreSQL (Compatible with Oracle), PolarDB-X, OceanBase, Oracle, and Dameng (DM)
NoSQL databases: Redis, MongoDB, Memcache, Cassandra, and Graph Database (GDB)
OLAP databases: AnalyticDB for MySQL and AnalyticDB for PostgreSQL
File and log storage: Object Storage Service (OSS) and Simple Log Service (SLS)
DMS supports Alibaba Cloud databases, databases from other cloud service providers, and self-managed databases. For details, see Supported database types and features.
Data operation and management
This layer covers three functional areas:
Data assets and security: Centralized management of data assets, security controls, and governance across databases, data warehouses, and data lakes.
Database design and development: Lock-free change, data import, schema synchronization, table consistency repair, database grouping, and data tracking.
Data transmission and processing: Data synchronization, data migration, change tracking, data integration, and orchestration and scheduling—all operable with SQL statements.
Data applications: Use DataService Studio and data analytics.
For a full feature list, see Features.
Commercial application
DMS supports the following business scenarios:
Data-based O&M
Multi-active disaster recovery
Next-day full data snapshot
Centralized data processing
Visual report
Multi-dimensional data analysis
Data Disaster Recovery
The Data Disaster Recovery module addresses the limitations of traditional backup approaches—long restoration times, high costs for large datasets, and risky manual processes.
Scenarios
Emergency disaster recovery
When a data security incident occurs, Data Disaster Recovery automatically retrieves backup sets and generates snapshots to create a sandbox instance. The sandbox maps directly to the source database instance and supports read/write operations without affecting source data.
Use the sandbox instance for:
Data restoration
Recovery drills
Development and testing
Query analysis
Emergency disaster recovery
For details, see DBS sandbox overview.
Performance specs:
Metric | Value |
Backup rate | More than 200 MB/s |
Maximum compression ratio | 7:1 |
Restoration time | 1 TB in 2 minutes |
Log capture | Real-time in-memory |
Point-in-time restoration | Supported |
Benefits:
High performance: Backup rate exceeds 200 MB/s with up to 7:1 compression.
Fast restoration: Restore 1 TB in 2 minutes, with support for point-in-time rollback.
Low storage costs: Data deduplication reduces backup storage costs.
Storage virtualization: Sandbox instances run independently on Alibaba Cloud. Read/write operations on the sandbox do not affect source databases.
Geo-redundancy
Data Disaster Recovery backs up data and logs to an OSS bucket in a different region from the source database. Configure a cross-region backup schedule to implement geo-redundancy. For details, see Overview.
Benefits:
High security: Data is transmitted over a virtual private cloud (VPC). The source database does not need Internet access enabled.
Flexibility: Configure daily default backups and monthly cross-region backups independently.
Fast data backup: Cross-region backup runs without requiring a leased line or Internet connection between regions.
Cross-cloud database backup
Run a backup schedule in Data Disaster Recovery to back up databases from third-party cloud platforms to Alibaba Cloud OSS. Source database endpoints are not exposed to the Internet—data is encrypted in transit.
Benefits:
High security and reliability: Encrypted transmission keeps source endpoints off the Internet.
Ease of use: A single backup and restoration platform handles databases across different cloud environments and reconciles different backup retention policies.
Database- and table-level restoration
Data Disaster Recovery reads backup data only from the specified table, rather than restoring the full instance. Combine this with incremental backup to restore data to any point in time. For details, see Restore data by database or table.
Benefits:
High efficiency: Targeted table reads reduce restoration time compared to full-instance restoration.
High flexibility: Combine with incremental backup for point-in-time restoration.
Backup integration for ISVs
For independent software vendors (ISVs) managing multiple database environments and backup scripts, Data Disaster Recovery provides a unified platform that consolidates cloud backups across databases.
Benefits:
Cutting-edge backup technologies: Reduces O&M workload by replacing multiple backup scripts with a single platform.
Cloud backup: Integrates backup data from multiple databases in the cloud.
Low O&M costs: Protect user data, prevent data loss, and reduce operational costs.