Controlling NCP slowdown

To alleviate the slowdown condition, you can take one or more of the following actions:

Select a buffer size that makes more efficient use of NCP buffers.
Install more memory in the communication controller.
Transfer inbound data to the CPU more quickly by decreasing DELAY.
Transfer outbound data to the network more quickly by increasing line speed.
Note: Many systems have much more outbound data than inbound. Therefore, increasing line speed can cause more data to be transferred out of the NCP through the line. This reduces buffer use.
If the NCP slowdown can be attributed to any given session or sessions, session pacing can be used to slow down the sessions causing the congestion.

NCP clears slowdown conditions by accepting outbound data from the CPU more slowly and by accepting data from the network more slowly. It stops polling and requests the host to stop writing data. Slowdowns are reduced when a device is not polled, if a PIU READ puts the NCP in a slowdown condition. The value specified for MAXDATA can control the speed with which data is placed on intermediate routing node (IRN) links.

The following actions can be taken to alleviate slowdown for local SNA controllers:

Increase the value of MAXBFRU. This allows the controller to send more data on every read channel program.
Use session pacing to control the amount of data VTAM® sends to the controller. Specify lower pacing values for primary-to-secondary pacing. This limits the number of PIUs that are sent to the controller by VTAM.