Your Browser is not longer supported

Please use Google Chrome, Mozilla Firefox or Microsoft Edge to view the page correctly
Loading...

{{viewport.spaceProperty.prod}}

Performance aspects for shared pubsets

&pagelevel(3)&pagelevel

Disks on shared pubsets are accessed directly by all sharers via the input/output path.

DMS meta-operations (e.g. Open, Close, Allocation) on the shared pubsets can be processed more effectively with the following measures:

  • Start SCA on the master processor of the shared pubset. The indicator START-SPEEDCAT in the MRSCAT entry should also be set on the slave processors to ensure that SCA is also started in the event of a master change.

  • In a shared pubset network containing processors with different levels of performance, it must be ensured that the DMS management requests of all sharers can be processed efficiently by the master processor. This requirement should also be fulfilled after a master change (see section “Generation of the shared pubset network”).

  • If there is a permanent and intensive DMS meta-operation load, the processor on which the greatest load of DMS meta-operations is expected should be used as the master processor, in order to avoid superfluous communication.

  • The MSCF connections should be established via the most reliable and fastest communication paths (BCAM routes).