Thanks a lot for both of your suggestions. It is an assumption that QOS will be on the network link, but this still doesn't account for a continually backed up host trying to send data. We are planning to utilize TCP as the transport. I figured TCP, QOS, and adjustments in the size of the log_fifo_size (num - The number of lines fitting to the output queue) might solve the issue. It is my understanding that the log_fifo_size, which does not have flow control, can be used to buffer data bound for destinations (either local disk or remote). I might be wrong in this understanding. I need the replay script to feed into the syslog data stream on the central syslog-ng server, since the central syslog-ng server feeds other applications. LeRoy