.
Many Oracle customers use the Standby Database as a disaster recovery solution. Yet should the primary site fail, any transactions in the current redo logs and any archived redo logs not yet transferred to the standby database are lost.
Other Oracle customers have been using the EMC SRDF product to remotely mirror (geo-mirror) the entire database to a failover site. Should the primary site fail, everything will be available at the remote location. Oracle has certified this approach when all volumes are mirrored synchronously. However each transaction results in multiple sends across the wire between the two sites. One for the tablespace and another for the redo log entry while a third change would be sent when that redo log was archived. This results in a higher cost for Telco lines and latency as distance increases. This approach works fairly well within the distance limits of ESCON links. Should the sites be farther than 60 kilometers apart, they would require two T3 communications lines. Not only are T3 lines extremely expensive, they also introduce a great deal of latency which badly hurts database performance.
A combination of Standby Database and SRDF synchronous solutions requires only the redo logs (online and archived) & control files to be Geo-Mirrored, drastically reducing the communication needs and the impact of latency. This provides a "no loss" disaster recovery solution applicable to a much broader market.
.
As shown in the following table, this configuration greatly reduces the amount of data that needs to be exchanged, as well as the impact of latency as distances become large. This makes geo-mirroring a viable solution with thousands of miles of separation between sites. Note shipping only the online redo logs provides even greater performance. However, this mode is not currently supported.
