OSS ensures availability from system design, product configuration, and other aspects.

Data durability

The following figure shows the data durability- and availability-related metrics of OSS.

Zone-disaster recovery

OSS provides zone-redundant storage to achieve zone-disaster recovery. In the zone-disaster recovery mode, objects are stored as replicas across three zones within a region. OSS regularly checks the integrity of the stored data. Business data can be processed even if data in an entire zone is destroyed. Data can be synchronized and copied across three zones in real time. Business can be failed over from a faulty zone to a normal zone, which you are not aware of.

Features of OSS zone-disaster recovery are as follows:
  • Data center-level disaster recovery capabilities: Data reliability reaches 99.9999999999%. When a data center becomes unavailable due to hardware faults or disasters, OSS still maintains high consistency. This capability ensures that business is not affected and no data is lost. You are not aware of the failover process. This feature can meet requirements of key business systems that require zero Recovery Point Objective (RPO) and Recovery Time Objective (RTO).
  • Higher SLA: OSS zone-redundant storage provides an SLA of 99.95%, which are five times higher than the SLA of the standard storage of data in a single zone.
  • One-click activation: OSS-based zone-redundant storage allows you to build cloud-based zone-disaster recovery capabilities with a single click. You can enable zone-redundant storage when you create a bucket. OSS uses a multi-replica mechanism to automatically store user data in three zones that are several kilometers away from each other within the same region.

OSS zone-redundant storage is available in China (Beijing), China (Shanghai), China (Hangzhou), and China (Shenzhen) on the China site (aliyun.com). This feature will be available in more regions in the future.

High reliability and stability

The three replicas are highly consistent and distributed across different zones. Data is automatically replicated when a fault occurs, as shown in the following figure.

Remote disaster recovery

The remote disaster recovery solution is mainly applicable to the following scenarios:
  • Compliance requirements: According to some industrial compliance regulations, a replica of the data must be stored at a site that is a certain distance from the original site where the data is stored.
  • Remote backup and disaster recovery: A replica of the data must be stored at a remote site in case of severe natural disasters, such as earthquakes and tsunamis.
  • Data replication: For business reasons, you may need to migrate data from one OSS IDC to another.

OSS supports cross-region replication for remote disaster recovery. It provides the following features:

  • In asynchronous replication mode, data latency is related to the amount of transmitted data and the transmission speed. Generally, the latency ranges from several minutes to several hours.
  • You can view the progress of historical data synchronization tasks in the console.
  • You can add, modify, or delete a synchronization task.
  • Filter rules: You can set the configuration to synchronize all objects in the source bucket or synchronize only objects with a specified prefix.

For more information, see Cross-region replication.