Tuning the behavior of prefetch queries can mitigate two possible performance issues: high System CPU utilization and inefficient network utilization.

The poll() mechanism employed in the OpenEdge client/server environment is known to have a high System CPU cost associated with it. In large-scale deployments, the cost of the poll() mechanism can overwhelm a system, preventing proper scaling. By managing the frequency of poll() calls, you can reduce System CPU utilization and increase user CPU utilization.

Additionally, using large network message buffers to make most efficient use of network bandwidth is often subverted when the remote client digests the received data faster than the server can populate a network message to its capacity. By managing the contents and behavior of network messages, you can more effectively manage your network bandwidth.

OpenEdge provides the ability to manage the contents and behavior of network messages and the frequency of calls to the poll() mechanism to mitigate these issues. Using tuning parameters, you can fine tune the behavior of message packing for prefetch queries as well as how often the poll() system call is invoked.