The Replication Time Control (RTC) feature provided by Object Storage Service (OSS) helps you meet compliance requirements or business requirements for cross-region replication (CRR). After the RTC feature is enabled, OSS replicates most of the objects that you uploaded to OSS within a few seconds and replicates 99.99% of objects within 10 minutes. In addition, RTC allows you to monitor replication metrics in real time.
Billing
After you enable the RTC feature, you are charged for data replication traffic that is generated for RTC-enabled data replication tasks. For more information, see RTC traffic fees.
For a source region and a destination region that you configured for a replication task, you can claim a Service Credit if the replication bandwidth limit and queries per second (QPS) limit are not reached, and the service-level agreement (SLA) is not met in the billing month. For more information, see Alibaba Cloud International Website Object Storage Service Service Level Agreement.
Limits
Regions
Regions in the Chinese mainland
RTC is available in the following regions: China (Hangzhou), China (Shanghai), China (Qingdao), China (Beijing), China (Zhangjiakou), and China (Shenzhen).
Regions outside the Chinese mainland
RTC is available in the following regions: US (Silicon Valley) and US (Virginia).
Bandwidth
Bandwidth limits between two regions
The bandwidth limit between two regions in the Chinese mainland is 10 Gbit/s. The bandwidth limit between two regions outside the Chinese mainland is 2 Gbit/s.
Bandwidth limits for one region
The bandwidth limit for a region in the Chinese mainland is 20 Gbit/s. The bandwidth limit for a region outside the Chinese mainland is 4 Gbit/s.
If you require higher bandwidth, contact technical support.
QPS
QPS limit for replication between two regions
The QPS limit between two regions in the Chinese mainland is 10,000. The QPS limit between two regions outside the Chinese mainland is 5,000.
QPS limit for replication for one region
The QPS limit for a region in the Chinese mainland is 20,000. The QPS limit for a region outside the Chinese mainland is 10,000.
ImportantThe preceding QPS limits apply to replication that uses non-sequential writes. The QPS limit for replication that uses sequential writes is 2,000. If QPS is too large, the latency increases. We recommend that you do not name the objects with sequential prefixes if you want to upload a large number of objects. For more information about how to change sequential prefixes to random prefixes, see OSS performance and scalability best practices.
Enable RTC
You can enable the RTC feature for new or existing CRR rule. You can enable the RTC feature for CRR rules to replicate all objects in a bucket, or to replicate objects with specified tags or whose names contain specified prefixes in a bucket.
If the RTC feature is enabled for a CRR rule in which the Replicate Historical Data parameter is set to No, the RTC feature takes effect approximately 15 minutes after the feature is enabled. If the RTC feature is enabled for a CRR rule in which the Replicate Historical Data parameter is set to Yes, the RTC feature takes effect approximately 1 hour after the historical data is replicated.
Monitor replication metrics
After the RTC feature is enabled, you can click RTC Metric Monitoring on the CRR tab in the OSS console to view replication metrics. For example, you can view the data replication latency, the amount of data remains to be replicated, and the number of operations to be replicated.
The following table describes the replication metrics.
Metric | Description |
RTC Replication Latency | The maximum data replication latency of the replication, that is, the recovery point objective (RPO) of the replication. For example, if the data replication latency is 1 minute, RTC ensures that all data written to the source bucket 1 minute ago is replicated to the destination bucket. When a disaster or an emergency occurs, you can recover the data to a point in time 1 minute ago. |
RTC Amount of Data to Replicate | The total size of data that remains to be replicated to the destination bucket by a replication rule at a specific point in time. For example, 800 MB of data is written to the source bucket at 08:00 and OSS replicates 600 MB of the data between 08:00 and 08:01. At 08:01, the amount of data that remains to be replicated is 200 MB. |
RTC Number of Operations to Replicate | The total number of operations that remains to be replicated by a replication rule to the destination bucket at a specific point in time. For example:
At 08:01, the number of operations that remains to be replicated is 10. |
RTC Write Bandwidth | The write bandwidth of the data to replicate from the source bucket by a replication rule. For example, the objects whose names contain the prefix1 and prefix2 prefixes are written to the source bucket at 08:00 at a total bandwidth of 800 Mbit/s, where:
If the replication rule replicates only the objects whose names contain the prefix1 prefix, the RTC write bandwidth is 200 Mbit/s. |
RTC Write QPS | The QPS of write operations to replicate from the source bucket by a replication rule. For example, the upload, modify, and delete operations are performed on the source bucket at 08:00.
If the replication rule replicates only the upload and modify operations performed on the source bucket, the RTC write QPS is 15. If the replication rule replicates all operations performed on the source bucket, the RTC write QPS is 18. Note The RTC write operations include the following operations:
|