Managing Latency Technical Brief

Managing Latency is One of the Most Critical Aspects of Data Replication Performance

The two most important types of latency are replication latency and application latency.

Replication Latency

• Replication latency is the elapsed time between when a data change (e.g., insert, update, delete) is committed to the source system database, and when it is applied to the target database.
• It is the time period when the change data is in the replication pipeline.
• It is affected by system utilization, process priorities, and network bandwidth.
• All data replication products suffer some amount of replication latency, regardless if they use asynchronous or synchronous technology.

This Time Period is Very Important to Businesses

It represents production change data that is not yet backed-up and may be lost in the event of an outage of the source or target system. The longer this time interval is, the more data may be lost.

Application Latency*

• When a data replication product uses synchronous replication, application latency can also occur.
• Application latency is the extra time spent after the source application calls transaction commit, before the commit completes.  It is caused by the extra work that the data replication engine must perform to safe-store (or apply) the change data on the target system before the source transaction is allowed to commit.
• Shadowbase Zero Data Loss (ZDL) implements synchronous replication by ensuring the source transaction’s data changes are safe-stored on the target system, before allowing the source transaction to commit.
• A consequence of this extra processing is that the time may be extended for the source application’s transaction commit to complete.
• This additional commit time is known as the application latency.
• Shadowbase ZDL thus ensures that no data will be lost in the event of an unplanned outage, but it may increase application latency as a consequence.

Controlling both replication and application latency is very important. This technical brief provides details and information regarding how to monitor, manage, and minimize both replication and application latency in Shadowbase environments that run on HPE NonStop systems.

*Application latency only applies when using the Shadowbase ZDL solution or in synchronous replication architectures.

Technical Brief:
Managing Latency in NonStop Systems