Search This Blog

Sunday, August 5, 2012

Different Modes of Data guard Protection


Introduction

  • When using an Oracle standby database for Business Continuity purposes there are 3 possible modes of operation for determining how the data is sent from the primary (the database currently being used to support the business queries) database to the standby (failover database to be used upon invocation of business continuity) database.
There are 2 synchronous modes and 1 asynchronous mode.

The various modes of operation of a standby are a real trade-off between data protection and impact on the primary systems.

 

Modes of Operation

The following table shows the options giving the benefits and drawbacks.

Mode of Operation
SYNC/
ASYNC
Benefits
Drawbacks
Maximum Protection
SYNC
Never any data loss, no data ever accepted on primary
without being on standby.
Slower response times on primary. Network disruption
between primary and secondary causes downtime of primary database.
Maximum Availability
SYNC
Updates must be on standby before accepted on primary,
network disruption means the primary continues running with updates sent to
standby when it is available again.
Slower response times on primary. Data loss possible if
network fails and then site fails.
Maximum Performance
ASYNC
No performance slowdown for applications on primary.
Data loss will happen

  • The customer is currently running in Maximum Performance mode, and thus the standby solution has minimal impact on the response times of the various applications.
  • There are 2 factors that would impact the application response if they moved to a higher level of data protection, the network latency and the network bandwidth. During normal operation, the response time of applications that make data changes would be impacted by the round trip time (rtt) to their standby
  • During periods of high volume activity, the bandwidth can become congested and this could lead to significant delay in response time.

 

Conclusion

  • If a higher level of data protection was deemed a business requirement it is advised that you have a significantly larger network bandwidth to the Business Continuity site.
  • During normal running we typically see a transport lag (of data between the sites) of around 10 seconds and this would be the typical amount of data we would lose in a disaster situation. During high peaks of activity this can increase.
  • In determining to increase the protection mode of the configuration there is a real trade off between how much you value your data (for this customer it is a reputation issue as much as a financial consideration) and the impact on your normal day-to-day performance.

1 comment: