Data Disaster Recovery provides a complete sandbox solution for MySQL and Oracle databases based on the copy data management (CDM) technology. You can use the sandbox feature to create a sandbox instance such as an ApsaraDB RDS emergency instance to restore data from a backup set. Read and write operations performed on sandbox instances do not affect source databases. You can perform various operations on sandbox instances, such as recovery drills, development and testing, query and analysis, and disaster recovery.
Background information
When you need to back up large amounts of data, traditional database backup services have the following shortcomings:
High storage costs: Large amounts of duplicate data are generated, and high storage costs are incurred. This is because full and incremental backups must be performed on a regular basis to ensure recovery point objectives (RPO).
High usage costs: You must restore data before it can be used. Large amounts of bandwidth, disk I/O resources, and CPU resources are required to restore large amounts of data. Restoration is costly and extends data recovery time objectives (RTO) to hours.
Data Disaster Recovery provides a complete sandbox solution for MySQL and Oracle databases based on the CDM technology.
MySQL sandbox instance: Data Disaster Recovery provides storage and computing resources for MySQL sandbox instances and hosts the service processes of the instances. You can remotely access the sandbox instances.
Oracle sandbox instance: Data Disaster Recovery provides storage resources for Oracle sandbox instances based on a remote file system and hosts the service processes of the instances. The computing resources of Oracle sandbox instances are provided by your Elastic Compute Service (ECS) instances.
The sandbox solution adopts various advanced backup technologies in the industry. The following figure shows the overall technical architecture of the solution.
Benefits
Physical backup of databases: accelerates data backup and restoration by backing up and restoring database files.
Data deduplication: significantly reduces the storage costs of backup data.
Storage snapshot and clone: allows backup data to be available in a short period of time. No data is copied.
Storage virtualization: provides sandbox instances of cloud resources without depending on your production instances. Read and write operations performed within sandbox instances do not affect source databases.
Scenarios
Development and testing based on online data (Dev or Ops)
Testing and prerelease are performed based on existing online data, and online data must be protected from being corrupted. The sandbox feature can provide replicas of full historical data and meet the requirements of rapid iteration, tests, and data reliability.
Online emergency disaster recovery
The sandbox feature allows the backup system to deliver an RPO or RTO within minutes and immediately restore online data. Compared with multi-active and hot backup, the sandbox feature also provides data rollback capabilities, which allows you to quickly roll back the system if dirty data is found.
Disaster recovery drill
Disaster recovery drills are performed on a regular basis based on the quick restoration capabilities and the pay-as-you-go billing method of sandbox instances. This is suitable for the scenarios in which high data reliability is required.
Ad hoc analysis of historical data
In data analysis scenarios, you must perform ad hoc queries on historical data based on your current requirements. The sandbox feature can provide replicas of full historical data for queries. Unlike solutions for offline historical data analysis based on big data platforms, the sandbox feature can provide SQL interfaces of original databases. This eliminates the need for developers to write offline analysis scripts and reduces storage costs by using snapshot storage.
Billing method
For more information, see Sandbox fees.