Managing Parallelism

Although the characteristics of package task and data flow transformation execution management are slightly different, each can be either serialized or asynchronous. Synchronous tasks use fewer system resources but might also take longer to process. Any task or data flow transformation can potentially spawn a separate execution thread. The maximum number of concurrent execution threads is controlled at the package level, using the MaxConcurrentExecutables property.

The default value of this property is set to -1. A property setting at this value causes the maximum number of executions to be two plus the number of processor cores in the computer. Set this to a different number to reduce or exceed the default. If SSIS runs on a dedicated server and you have a lot of operations that run in parallel, you will likely want to increase this setting if some of the operations do a lot of throttling or waiting for external systems to reply. Alternatively, if you do not have a dedicated SSIS machine and your data integration application runs alongside several other applications, you might need to reduce this setting to avoid resource conflicts.

As you design packages for parallelism, you will need to decide whether to run some or all of the operations in the package in parallel. As with buffer-sizing decisions, decisions about parallelism are best made when you take into account available system resources on your production server.