You can reduce the number of socket descriptors to be used by a single scheduler daemon by distributing the client UAPs to be connected to scheduler daemons.

The figure below shows an example of a system configuration that solves the deficiency of socket descriptors.

Figure C-8 Example of a system configuration that solves the deficiency of socket descriptors

You can reduce the number of connection establishment requests to be held in the wait queue for the listen system call by distributing the client UAPs to be connected to scheduler daemons.

The figure below shows an example of a system configuration that solves errors with the connect system call.

Figure C-9 Example of a system configuration that solves errors with the connect system call

You can effectively use a network having a high line speed by separating the scheduler daemon for processing client UAPs on a network having a high line speed from the scheduler daemon for processing client UAPs on a network having a low line speed.

The figure below shows an example of a system that effectively uses a network having a high line speed.

Figure C-10 Example of a system that effectively uses a network having a high line speed

By using the multi-scheduler facility and localizing scheduler daemons whose message receive processing is interrupted, you can schedule messages without delaying other service request messages.

The figure below shows an example of a system in which service request messages are not interrupted.

Figure C-11 Example of a system in which service request messages are not interrupted

You can use the multi-scheduler facility to increase the number of processing threads that can be executed at one time. This prevents temporary communication errors and time-out caused by a deficiency of processing threads.

The figure below shows an example of a system with an increased number of simultaneously executable processing threads.

Figure C-12 Example of a system with an increased number of simultaneously executable processing threads