Summary of Replication Options
The following table summarizes different aspects of each implementation.
- What can be replicated?
What artifacts can replicate this implementation: any file, a specific type of storage, etc.
- How is the replica performed?
Whether the replica is automatically done by OCI, or if you need to create and run a script.
- Connectivity requirements
Indicates the network connectivity requirements specific to this replica implementation.
- Additional SW/HW requirements
Whether the replica requires additional components like a central host, central storage, or additional software.
- Staging storage requirements
Whether the replica implementation needs a staging storage or not.
- Staging storage setup and management complexity
If a staging storage is needed, how complex is it to set up and manage it.
- Switchover and failover complexity
The complexity (in terms of low, medium, high) of the procedure for making the contents available and usable in the secondary site. This analysis doesn’t include the switchover and failover tasks of other components of the system (for example, database switchover, DNS changes, stop/start processes).
- RTO in switchover
Refers to the time needed to make the replicated contents available and usable in the secondary nodes, which causes an increase of the downtime during a switchover. This doesn’t include the time spent in the last replication; it just refers to the time spent in actions like attaching or mounting the storage. It only refers to the tasks related to the replicated storage; it does not include the downtime of the tasks for switching over other components of the system (like database switchover, stop/start processes, etc.).
- RTO in failover
The time needed to make the replicated contents available and usable in the secondary, which causes an increment of the downtime during a failover. This doesn’t include the time spent in the last replication; it just refers to the time spent in actions like attaching or mounting the storage. It only refers to the tasks related to the replicated storage; it does not include the downtime of the tasks for failing over other components of the system (like database failover, stop/start processes, and so on).
- TCO (Total Cost Ownership)
The complete cost of owning the system over its entire lifespan, including not just the licensing price, but also all associated costs like maintenance, operation, etc. This point is separated into the billing costs and the operational costs.
- Additional supportability restrictions
List any information about restrictions to support this method.
| Mid-Tier Replica Implementation | OCI Block Volumes Replication | OCI File Storage Replication | Rsync Peer-to-Peer Replication
|
Rsync with Central Staging Replication
|
Oracle Database File System (DBFS) Replication |
|---|---|---|---|---|---|
| What can be replicated? | The Block Volumes only
(BV does not allow sharing across nodes like a shared storage). |
The OCI File Storage file systems only, which is shared storage (normally containing artifacts shared across multiple nodes). | Any file artifacts, regardless of the underlying storage. | Any file artifacts, regardless of the underlying storage. | Any file artifacts, regardless of the underlying storage. |
| How is the replication performed? | OCI native storage replication (Block Volumes). | OCI native storage replication (File Storage) | Scripts | Scripts | Scripts |
| Connectivity requirements | No explicit connectivity is required. Replication happens through OCI’s backbone connections. | No explicit connectivity is required. Replication happens through OCI’s backbone connections. | SSH connectivity between the mid-tier peer nodes. | SSH connectivity from the bastion host to all the mid-tier hosts (primary and secondary). | SQL connectivity from the mid-tier hosts to the remote database may be required for role conversions. |
| Additional SW/HW requirements | N/A | N/A | N/A | Yes, a bastion host with a central storage. | Yes, an Oracle Database is required, and the DB client software must be installed in the mid-tier hosts.
Additional capacity at the storage and at processing levels may be required depending on the quantity and frequency of data. |
| Staging storage requirements | N/A | N/A | N/A | Yes, in the bastion. | Yes, the DBFS mount. |
| Staging Storage setup and management complexity | N/A | N/A | N/A | Low | Medium-High |
| Switchover and Failover complexity | Medium (Low using OCI Full Stack DR service) | Medium (Low using OCI Full Stack DR service) | Low | Low | Low |
| RTO in Switchover |
0 The tasks to activate, attach, and mount can be done earlier, so they do not affect RTO. |
0 The tasks to clone, export, and mount can be done earlier, so they do not affect RTO. |
0 The storage remains mounted in secondary; no previous action needed. |
0 The storage remains mounted in secondary; no previous action needed. |
0 The storage remains mounted in secondary; no previous action needed. |
| RTO in Failover |
5-10 minutes It can vary depending on the number of Bock Volumes and if the actions are performed in parallel. |
5-10 min It can vary depending on the number of FS and if the actions are performed in parallel. |
0 | 0 | 0 |
| RPO |
<30 minutes It can vary depending on the change rate of data in the source volume. |
15 min (minimum replication interval) It can vary depending on the change rate of data in the source volume. |
Depends on the frequency of replica script execution. |
Depends on the frequency of replica script execution. |
Depends on the frequency of replica script execution. |
| TCO Billing cost (related to the replica) |
Low Cost is network usage plus mid-tier storage. |
Low Cost is network usage plus mid-tier storage. |
Low Cost is network usage plus mid-tier storage. |
Low-Medium Cost is network usage plus mid-tier storage plus staging storage. |
Low-Medium Cost is network usage plus mid-tier storage plus staging storage in the database. |
| TCO Operational costs (maintenance, operation) |
Medium / Low * Managed by OCI infrastructure (*) When using Full Stack DR service for role changes |
Low / Very Low* Managed by OCI infrastructure (*) When using Full Stack DR service for role changes. Since FS is shareable, the expected number of elements to switchover is lower than in BV replica. |
Medium Managed by the customer. |
Medium Managed by the customer. |
Medium-High Managed by the customer, and requires different teams and knowledge (DBA). |
| Additional supportability restrictions | N/A | N/A | N/A | N/A | Not applicable if using an Autonomous Database. |