Previous Topic: Virtual Control Files

Next Topic: Tuning Control File Access Rates

Control File Mediums and Performance

The hardware medium that is used to store the CA MIM Physical Control File (DASD or a coupling facility) or transmit the CA MIM virtual control file (CTC or XCF) has a direct impact on the ability of CA MIM to process work efficiently. The faster CA MIM can read and write transactions to the CA MIM control file, the faster the CA MIM service requestors are dispatched and, in turn, greater system throughput is achieved.

One of the measurements used to gauge control file performance is the Average Cycle Time displayed with the CA MIM DISPLAY IO command. The Average Cycle Time reflects how long a control file cycle takes to complete. Control file cycle time begins at the moment the control file reserve is requested, includes the time it takes to obtain the reserve and read and write transaction data to the control file, and ends when the control file is released. This performance measurement is accurate for both physical and virtual control files.

The following table shows reasonable Average Cycle Times for various CA MIM control file mediums and architectures. This table is intended to show comparative rates between the various mediums and architectures.

Average Cycle Time

Control File Type

Control File Medium

CA MIM address spaces

25 ms

physical

ESCON DASD

all

15 ms

physical

FICON DASD

all

10 ms

virtual

CTC

client

10 ms

virtual

XCF

client

6 ms

virtual

CTC

master

6 ms

virtual

XCF

master

4 ms

physical

XES

all

Due to numerous environmental factors, the Average Cycle Times in a particular enterprise may be slightly higher or lower than those values shown here. You can see from this table that the type of medium used to store or transmit the CA MIM control file has a direct impact on the performance of the product. Again, the faster the control file cycle, the faster CA MIM transactions are processed, and the better the complex-wide enqueue throughput.