Avoid and remove bottlenecks

By running multiple tools simultaneously, Data Management reduces the amount of time needed to perform a given task, provided that there is no "bottleneck" in the process flow. A bottleneck can be one of several things:

• Split-merge bottlenecks: Anything that causes a part of a split record stream to wait until the other part is finished, for example a left-to-right merge. Because Data Management will convert the Concatenate tools of older projects into the equivalent Left-to-right Merge, this is a common issue when converting old projects.

• Choke points: Anything that causes all other processing to wait for a single tool to finish, such as a sort.

• Slow tools: Any single tool that is significantly slower than any other (for example, Address Standardization).

While bottlenecks are hard to completely eliminate, it is possible to reduce their impact by replacing suboptimal data flow patterns with better ones.