SCA (Speed Catalog Access) is a recommended software product for accelerating catalog services.In searches for file entries and job variable entries, SCA replaces sequential searches for catalog blocks with direct accessing.
This is accomplished using two tables. One table provides information about free space in the catalog blocks, and a second (reference table) assigns the logical block number of the catalog block to the file names or job variable names.
Both tables are created by the SCA task.
SCA can only be called for SF pubsets. In the case of SM pubsets, accelerated access to catalog entries is already implemented internally.
Two versions of SCA are available: with and without task switching. The use of SCA without task switching is recommended.
SCA without task switching
The reference table is created in class 4 memory and is thus available to all tasks. All SCA routines (except for table creation) run under the accessing user task. There is no overhead for task switching, and the overhead for serialization measures is minimal.
SCA with task switching
The reference table is managed by the SCA task in its class 5 memory.
Each catalog service request from a task results in a task switch.
You can store the variants that are to be used for a pubset or specify whether SCA is to be started automatically (standard for BS2000/OSD V8.0 and higher) when the pubset is imported in the MASTER catalog MRSCAT (START-SPEEDCAT operand in /ADD- and /MASTER-CATALOG-ENTRY).
The “SCA without task switching” version may be expected to require approx. 1 to 2 % less CPU time, but somewhat more class 4 memory than does the “SCA with task switching” version.
Automatic startup of SCA
If SCA is installed SCA is started in *SPEEDCAT-TASK mode when a pubset is imported in order to guarantee high-performance catalog access with SF pubsets.
The default value START-SPEEDCAT=*AUTOMATIC is therefore used in the ADD-MASTER-CATALOG-ENTRY command.
Advantages of using SCA
Increased total throughput rate for the system when the workload involves an intensive use of the catalog.
The reduction in the number of physical accesses of TSOSCAT per logical catalog call (OPEN, CLOSE, ...) results in a lowering of the system workload and a marked reduction in the catalog access time.
In addition, the burden on the disk drives containing the catalog files is reduced. If SF pubsets are used, these are generally the xxxx.0 volumes. As far as the whole system is concerned, this means that the throughput of the accessing user tasks can be increased.
Runtime for programs with a high frequency of catalog access (e.g. ARCHIVE) as well as response times for commands which need to access the catalog regularly (e.g. /CREATE-FILE, /MODIFY-FILE-ATTRIBUTES, /IMPORT-FILE) can be improved considerably.SCA is indispensable for fast systems operating in interactive mode. Its use leads to a marked increase in throughput.
Greater flexibility in computer center organization
The use of SCA makes one of the previous organizational methods of improving catalog performance redundant. It is therefore no longer necessary to place frequently used catalog entries at the beginning of a catalog block chain.
Full compatibility
SCA does not change the structure of the catalog. As a result, both the user interfaces and the internal system interfaces to the catalog management system are retained.
SCA can be started and/or terminated for different catalogs simultaneously, either automatically, when the pubset is imported (determined by the MRSCAT entry), or by procedure call (/ENTER-JOB FROM-FILE=SPEEDCAT.ENTER.START) after the pubset has been imported.