Zane Appel

Comparison between Hitachi Replication and Oracle Replication for Oracle Database

Blog Post created by Zane Appel on Apr 8, 2014

Most Oracle Database Administrators are familiar with the different replication products specifically for the Oracle Database. This is a comparioson between those replication products and the main replication products available on Hitachi Data Systems Storage arrays.

 

 

Hitachi True CopyHitachi Universal ReplicatorData GuardStreamsGolden Gate / SharePlex
Description

Hitachi Data Systems Storage level synchronous

replication

Hitachi Data Systems Storage level asynchronous

replication

Oracle database replication. Automated log shipping.Oracle database data and object sharing. Not replication per se but you can write your own replication using it.Oracle database data sharing platform. Primarily an enterprise application integration platform but can be used for replication as well.
Shareplex uses JMS while GG can be configured to use it.
Synch/AsynchSynchronousAsynchronousTypically Asynchronous. Synchronous is an option.AsynchronousAsynchronous
RPOZero RPO.
Zero data loss when fence=data.

 

Near zero to very low when fence=never.
Low RPO.
Depends only upon the amount of transactions and the capacity of the communication channel.
Asynchronous: Medium RPO.

 

Synchronous: Zero RPO.

 

Configurable. Typically will transmit each redo log as it is archived so depends upon how much work is being done.
Low to High RPO.
Depends upon how it is written. "In flight" data is lost.
Low to High RPO.
Depends upon how it is written. "In flight" data is lost.
However: If it uses JMS (Shareplex by default) the data is not actually lost but may not be available for recovery.
RTOLow RTO.
The database can be brought online after the replicated LUNs are made read / writeable. (Horctakeover)
Low RTO.
The database can be brought online after the replicated LUNs are made read / writeable. (Horctakeover)
Low RTO.
Warm Standby database can typically be used after preparation.
Low to Medium RTO.
Depends how it is written.
Typically Warm Standby.
Low to Medium RTO.
Depends how it is written.
Typically Warm Standby.
Latency Overhead Noticeable at longer distances.Negligible.Asynchronous: Negligible.
Synchronous: Noticeable at longer distances.
Medium to significant.
Depends how it is written.
Low to medium.
MaintenanceLow.
Few moving parts, none at database level.
Fully supported by Hitachi Data Systems.
Low.
Few moving parts, none at database level.
Fully supported by Hitachi Data Systems.
Medium.
Failures are common although it is much improved.
Supported by Oracle.
Common skillset for Oracle DBAs.

Potentially very high.
Since it is done manually there is no support, especially if author leaves the company.

Less common skillset for Oracle DBAs.

Not Sure, probably medium.
Fully supported by Oracle / Dell respectively.

Less common skillset for Oracle DBAs.

Outcomes