Before CA MICS can decide if a job has met its turnaround
target, the actual turnaround time must be computed.
Turnaround times are calculated as the sum of the durations
your site chooses to include in the calculations.
The four TURNTIME statements define how turnaround time is to
be computed for standard jobs submitted locally, standard
jobs submitted remotely, deadline jobs submitted locally, and
deadline jobs submitted remotely. (Standard jobs have a
JOBGROUP value between 1 and 149; deadline jobs have a
JOBGROUP value between 150 and 179.)
EXAMPLE
The following example illustrates a turnaround time
definition.
TURNTIME STANDARD LOCAL 1 2 3 7
TURNTIME STANDARD REMOTE 1 2 3
TURNTIME DEADLINE LOCAL 1 8
TURNTIME DEADLINE REMOTE 1 6
CODING THE TURNTIME STATEMENTS
As with most prefix.MICS.PARMS members, the format of
statements is free-form, but positional. Blank statements
are allowed. Comments are coded by beginning a statement
with an asterisk (*). You should modify the sample JOBGROUP
member's TURNTIME statements distributed with the product
rather than coding them from scratch.
Note that all four statements are required, even if you have
no deadline job groups or you are running JES3. (CA MICS can
only tell the difference between local and remote submission
for systems with JES2; it treats all jobs as local in systems
with JES3.)
The available parameters meet the needs of almost all sites;
if you have a special requirement, it can be met by the
TURNRTE turnaround time calculation exit (see Section 7.3.4).
TURNTIME STANDARD STATEMENTS: JOBGROUPS 1-149
The values specified after TURNTIME STANDARD correspond to
the duration of phases in the life of the job. Specify the
numbers that designate only those durations you want to be
summed in the computation. The values, their definitions,
and the variable names that hold the corresponding times are
as follows:
1 - Reader Active Time: Time required to read the job
into the system (JOBRDRTM).
2 - Converter Time: Time required to process the job's
JCL and convert it for JES processing (JOBCVTTM).
3 - Input Queue Time: Time that the job resided in the
input queue prior to initiator selection (JOBINQTM).
4 - Total Enqueue Time: Time that the job was initiated
but idle due to delays related to data set enqueue
or time spent waiting for V=R memory (JOBENQTM).
5 - Total Allocation Time: Time that the job was
initiated and idle due to delays related to device
allocation and/or volume mounting (JOBALCTM). Note
that JOBALCTM is the summation of step allocation
time for all steps in the job.
6 - Execution Time: Time that the job's program steps
were loaded and in some phase of execution status
(JOBEXCTM).
7 - Printer Queue Time: Time that the job's spooled
output resided in the output queue from job
termination until the time that the first data set
starts printing (JOBPRQTM).
8 - Printer Time: Time that the job's first spooled
output started printing until the time the job was
purged from the system (JOBPRNTM). Note that
printing is not necessarily going on all during this
interval, especially for jobs with multiple SYSOUT
data sets. In many cases, this value can be hours or
even days.
9 - Job Elapsed Time: Time that the job was selected for
initiation until job terminated, regardless of any
events that may have taken place, such as contention,
duplicate jobs, abends, restarts, and so on.
In the example above, turnaround time is computed as the sum
of input read time, conversion time, input queue time, and
printer queue time for STANDARD jobs submitted from a local
reader. STANDARD jobs submitted remotely are similar, except
that printer queue time is not counted as part of their
turnaround time.
Note: JOBESPTM (9) CANNOT be specified in a TURNTIME STANDARD
statement with JOBENQTM (4) or JOBALCTM (5) or JOBEXCTM (6).
They are mutually exclusive.
TURNTIME DEADLINE STATEMENTS: JOBGROUPS 150-179
The values specified in the TURNTIME DEADLINE statements
correspond to a starting and an ending timestamp,
respectively. For deadline jobs, CA MICS is not concerned
with the length of time between these two timestamps.
Rather, it wants to compare these timestamps to the "in by"
and "out by" times associated with the job group. The
allowable starting timestamps and their meanings are:
1 - The time of job submission (RDRTS).
2 - The time conversion ended (beginning of wait in the
input job queue).
3 - The time the job was selected for initiation (end of
wait in input queue/beginning of z/OS execution
(STARTTS)).
The allowable ending timestamps and their meanings are:
3 - The time the job was selected for initiation (end of
wait in input queue/beginning of z/OS execution
(STARTTS)). Note that 3 cannot be designated as
both the start and end timestamp.
6 - The time the job terminated (end of z/OS execution
(ENDTS)).
7 - The time the first line of output was printed for
the job.
8 - The time the job was purged (JOBPURTS).
Note: Some of these timestamps are not carried in the
CA MICS database. They are available at the time that
deadline turnaround time is computed, and can be
reconstructed, if necessary, from RDRTS and the durations of
the variables listed in the documentation of STANDARD job
turnaround.
In the example in the prior section, DEADLINE jobs submitted
locally were considered to have become available to the
system when they were read in, and to have ended when they
were purged. For DEADLINE jobs submitted remotely, the start
time is the same, but the end time is when the job finished
z/OS execution.
FACTORS TO CONSIDER WHEN DEFINING TURNAROUND TIME
There is no universally acceptable definition of job
turnaround time. Generally, users do not hold the computer
center responsible for parts of the job's life over which it
has no control, and do hold it responsible for those phases
of processing over which it does have control; turnaround
definitions reflect this responsibility.
For example, for standard jobs, turnaround is the sum of the
lengths of various phases of the jobs' lives in the system.
If the site has no control over the resource consumption of
submitted standard jobs, then it cannot be held responsible
for the z/OS execution phase of those jobs' lives; they may
run a minute or they may run an hour. On the other hand,
most shops can, by proper scheduling, control how long a job
must wait to begin z/OS execution, and their turnaround time
definitions would include this time.
For deadline jobs, the same considerations apply, but are
expressed somewhat differently since such jobs have their
turnaround expressed in terms of just a starting time and an
ending time. In this case, the ending time would be set to
the point at which the job entered z/OS execution to exclude
the run time.
The source of the job also has a bearing on those phases of
job life which can be controlled by the computer center, and
therefore should be included in the turnaround time
computation. For example, for jobs submitted from a remote
location that is staffed by a user department, the data
center cannot control when the completed jobs will be allowed
to start printing. Thus print queue time would be excluded
from the turnaround computation for such jobs.
On the other hand, the data center may decide to include the
print queue time for jobs submitted and printed locally.
Doing so would be unwise if many jobs were submitted from TSO
with their output held for viewing at the terminal. Such
jobs may stay in what appears to be the print queue for days
and then be purged without ever printing when the user
cancels them from the terminal.
+--------------------------------------------------------------------------+ | INSTALLATION PREPARATION WORKSHEET: Turnaround Time Definition | | | | PARMS Library Member is JOBGROUP | |Reference: Section 7.3.2.2 CA MICS Batch and Operations Analyzer Guide | +--------------------------------------------------------------------------+ | | | Use the numerals listed below to designate the components of turnaround | | time for STANDARD JOBGROUPs. The name of the data element in the | | BATJOB file that holds the value is listed in parentheses. | | | | 1 - Reader Active Time (JOBRDRTM) | | 2 - Converter Time (JOBCVTTM) | | 3 - Input Queue Time (JOBINQTM) | | 4 - Total Enqueue Time (JOBENQTM) | | 5 - Total Allocation Time (JOBALCTM) | | 6 - Execution Time (JOBEXCTM) | | 7 - Printer Queue Time (JOBPRQTM) | | 8 - Printer Time (JOBPRNTM) | | 9 - Job Elapsed Time (JOBESPTM) | | | | | | TURNTIME STANDARD LOCAL _ _ _ _ _ _ _ _ | | TURNTIME STANDARD REMOTE _ _ _ _ _ _ _ _ | |__________________________________________________________________________| | | | Define DEADLINE JOBGROUPS using the comments below to designate | | starting timestamps: | | | | 1 - Time of job submission (RDRTS). | | 2 - Time conversion ended (beginning of wait in | | the input job queue). | | 3 - Time the job was selected for initiation (end of wait in the | | input queue/beginning of MVS execution (STARTTS)). | | | | Allowable ending timestamp numbers are: | | | | 3 - Time the job was selected for initiation (end of wait in the | | input queue/beginning of MVS execution (STARTTS)). | | 6 - Time the job terminated (end of MVS execution (ENDTS)). | | 7 - Time the first line of output was printed for the job. | | 8 - Time the job was purged (JOBPURTS). | | | | The specification of "3" as both the start and end timestamp | | designation is not allowed. | | | | TURNTIME DEADLINE LOCAL _ _ | | TURNTIME DEADLINE REMOTE _ _ | | | +--------------------------------------------------------------------------+ | ....5...10...15...20...25...30...35...40...45...50...55...60...65...70.. | +--------------------------------------------------------------------------+
Figure 7-6. TURNTIME Statement Worksheet for JOBGROUP Member
|
Copyright © 2014 CA.
All rights reserved.
|
|