11. WORKLOAD CHARACTERIZATION › 11.6 Case Studies › 11.6.1 Batch Job Class › 11.6.1.2 Control Parameters
11.6.1.2 Control Parameters
Figure 11-18 illustrates the Workload Characterization
screen.
/------------------------- Workload Characterization ------------------------\
|Command ===> |
| |
|Composing CA MICS Inquiry: CLASS - Batch Job Class Case Study |
| |
|CA MICS file ===> JOB - |
|Time-span: DETAIL Cycle(s) ===> 01 - 07|
|Data base ===> A _ _ _ _ |
|Allow Run-Time Override of Data Base ===> Y (Y/N) and/or Cycle(s) ===> Y (Y/N)|
| |
|Analysis based on ===> JOBMXNTA JOBTCBTM JOBNLR__ ________ ________ |
| ________ ________ ________ ________ ________ |
|Report for each cluster ===> JOBTCBTM JOBEDASD INITMIN JOBNLR ________ |
| ________ ________ ________ ________ ________ |
| |
|Data Selection: Allow run-time override|
| Date ===> _______ - _______ _______ - _______ ===> N (Y/N)|
| Hour ===> __ - __ __ - __ __ - __ ===> N (Y/N)|
| Zone ===> _ _ _ _ _ _ _ _ _ ===> N (Y/N)|
| SYSID ===> ____ ____ ____ ____ ____ ____ ===> N (Y/N)|
| |
| |
|Specify extended options ===> Y (Y/N) |
\------------------------------------------------------------------------------/
Figure 11-18. Workload Characterization Screen Control Parameters
The Workload Characterization screen lists the following
fields:
CA MICS file: The CA MICS file to be used for workload
characterization. (For this example, we
specified the CA MICS BATJOB File for a job
class study.)
Cycles: The number of cycles to be specified. (For
example, we specified seven cycles to analyze a
full week of data. In an actual study, analyze
as much data as possible before you initiate a
new job class structure.)
Database: The single-character database identifier for
this file at your site.
Analysis The CA MICS data element feature names (that
based on: is, tape drives, CPU time, and print lines) to
be used in the clustering procedure. (As
discussed in Section 11.2.4, you should specify
as few features as possible.)
CPU time, DASD EXCPs, initiator-occupied
minutes (that is, elapsed time), and print
lines are specified for reporting purposes.
(Because initiator-occupied minutes is not a
standard element in the CA MICS database, it is
calculated using the user-defined macro DRVMAC
that is specified on another screen).
/--------------------------- Extended Options Menu --------------------------\
|Command ===> |
| |
|Inquiry Step: Workload Characterization |
| |
| |
|Specify global data selection exit ===> N (Y/N/R) |
|Specify non-CA MICS data element derivation exit===> Y (Y/N/R) |
| |
|Specify alternate input file ===> N (Y/N/R) |
|Save selected input observations ===> N (Y/N/R) |
|Save cluster output observations ===> Y (Y/N/R) |
| |
|Specify statistical analysis parameters ===> Y (Y/N/R) |
|Define report description element names ===> Y (Y/N/R) |
| |
| |
| |
| |
| |
| |
| |
| |
\------------------------------------------------------------------------------/
Figure 11-19. Workload Characterization Extended Options Menu
You may select additional option specifications on the
Extended Options Menu. Figure 11-19 shows an example in
which we specified four of the extended options.
/------------------------ Report Description Elements ----------- ROW 1 OF 2\
|Command ===> Scroll ===> CSR |
| |
|Inquiry Step: Workload Characterization |
| |
|Line cmds: I Insert D Delete M Move |
| |
|Cmd Element Data element label (long name) Warning |
| - -------- ---------------------------------------- -------- |
| _ JOB |
| _ JOBCLASS |
|****************************** BOTTOM OF DATA *******************************|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
\------------------------------------------------------------------------------/
Figure 11-20. Report Description Elements Screen
Since we wished to create an optional external report data
set during the study, we also specified two descriptive
variables. We selected these variables, JOB (job name) and
JOBCLASS (job class), so that the results of the cluster
analysis could be compared with the current job class
structure and the assignment of particular jobs could be
determined. These data elements are specified on the Report
Description Elements screen shown in Figure 11-20.
/------------------------------- SAS Statements -----------------------------\
|Command ===> Scroll ===> CSR |
| |
|Inquiry Step: Workload Characterization |
| |
|Line Cmds: I Insert D Delete R Repeat M Move C Copy |
| |
|Cmd SAS Statements for: Specifying data element derivation |
| - ------------------------------------------------------------------------ |
| _ - %MACRO DRVMAC; |
| _ INITMIN=(ENDTS-STARTTS)/60; |
| _ - %MEND DRVMAC; |
|****************************** BOTTOM OF DATA *******************************|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
\------------------------------------------------------------------------------/
Figure 11-21. User-Defined Data Element Derivation Macro
The SAS statements that compute the user-defined data
elements are placed in the DRVMAC macro shown in Figure
11-21.
/---------------------- Statistical Analysis Parameters --------- ROW 1 OF 1\
|Command ===> Scroll ===> CSR |
| |
|Inquiry Step: Workload Characterization |
| |
|Sample size ===> 2000 (nnnnnn) |
|Observation trimming (percent) ===> 97.5 (nn.n) |
|Objective radius ===> 1.5 (1.0-2.0) |
| |
|Delete observations if the following elements contain SAS missing values or 0:|
| |
|Cmd Element Data element label (long name) Warning |
| - -------- ---------------------------------------- -------- |
| _ JOBTCBTM |
|****************************** BOTTOM OF DATA *******************************|
| |
| |
| |
| |
| |
| |
| |
| |
\------------------------------------------------------------------------------/
Figure 11-22. Delete Observations
We excluded jobs that used no CPU time from the study, since
they did not execute due to a JCL error, or some other
abnormal condition. To exclude jobs with no CPU usage,
specify the data element name, JOBTCBTM, on the Statistical
Analysis Parameters screen shown in Figure 11-22.
/--------------------- Save Cluster Output Observations ---------------------\
|Command ===> |
| |
|Inquiry Step: Workload Characterization |
| |
|SAS file name ===> ALLOBS |
| |
|SAS data set name ===> 'your.dataset.name' |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
\------------------------------------------------------------------------------/
Figure 11-23. External Report Output File Definition
In this case study we wanted to create the SAS data set name,
ALLOBS, that contains the cluster analysis output. To do
this, we specified ALLOBS on the External Report Output File
Definition screen shown in Figure 11-23.
/----------------------------- Inquiry Processing ----------------------------\
|Command ===> |
| |
| |
| |
| Active MICF Inquiry: CLASS - Batch Job Class Case Study |
| Author: Charles Frost |
| |
| |
| Specify Inquiry processing ===> EXECUTE (EXECUTE/MODIFY/SAVE/CANCEL) |
| EXECUTE - Execute the inquiry. |
| MODIFY - Redisplay the inquiry for modification. |
| SAVE - Exit from inquiry composition and save the inquiry. |
| CANCEL - Exit from inquiry composition without saving the inquiry. |
| |
| |
|Instructions: |
| |
| Specify desired option and press ENTER key. |
| |
| |
| |
| |
\------------------------------------------------------------------------------/
Figure 11-24. Inquiry Processing
After a workload characterization analysis inquiry is
created, the Inquiry Processing screen shown in Figure 11-24
is displayed. From this screen, you can execute the newly
created inquiry, modify it, save it without executing it, or
cancel the parameter specifications you just entered.
Copyright © 2014 CA.
All rights reserved.
|
|