5. DATABASE FILES › 5.2 SNT Information Area › 5.2.6 NPM System (PLU) Activity File (SNTPSY) › 5.2.6.3 Usage Considerations

5.2.6.3 Usage Considerations

This section identifies the special considerations or
techniques related to using the SNTPSY file, followed by
several retrieval examples.


In the examples, a SAS macro variable is used to specify the
DDname part of the CA MICS file name. These macro variables
are a standard part of CA MICS and are available for all
files. The macro variable name has the form &diiit, where d
is the database identifier, iii is the information area
name, and t is the timespan.  For the examples, a database
identifier of P is used.  The identifier is installation
dependent, so you should find out what the identifiers are at
your installation.

Special Considerations/Techniques

1. Care must be exercised in using the special date and time
   data elements contained in each CA MICS file. As the
   file's granularity increases in higher timespans, certain
   fields lose significance and should not be used in those
   cases.

   o HOUR should not be used in MONTHS.
   o DAY and DAYNAME should not be used in WEEKS or MONTHS.
   o WEEK should not be used in MONTHS.

2. The following data elements are user-defined.  The value
   in each depends on user-selected options or user-supplied
   SAS code:

   o NETWRKID - Network Identifier.  A one to eight-character
     descriptive name for each network represented in the
     data.  NETWRKID is set in the Network Identifier Exit
     Routine, SNTNIDRT, discussed in Chapter 7.  For NPM,
     NETWRKID may be set to the value of PLU Network
     Identifier (PLUNETID).

   o TMOFFSET - Time offset applied to the data in the file.
     The user can associate a time offset with each unique
     NETWRKID/SYSID combination.  This offset is used to
     adjust data times when data from multiple time zones is
     brought together in the CA MICS database. To determine
     the original time at which the data was recorded,
     subtract TMOFFSET from the value.  Time offset is
     assigned in the SNTOPS member of prefix.MICS.PARMS,
     discussed in Chapter 7.

3. The following data elements lose significance once
   summarization has been performed.  These data elements
   should be referenced only when the last observed value in
   a summarization interval is significant.

       NPMHIHST - High Threshold Time, For Host
       NPMHINET - High Threshold Time, For Network
       NPMHITOT - High Threshold Time, Total
       NPMLOHST - Low Threshold Time, For Host
       NPMLONET - Low Threshold Time, For Network
       NPMLOTOT - Low Threshold Time, Total
       NPMRVAL1 - Response Distribution Limits 1
       NPMRVAL2 - Response Distribution Limits 2
       NPMRVAL3 - Response Distribution Limits 3
       NPMRVAL4 - Response Distribution Limits 4
       SLU      - Secondary Logical Unit
       SLULINK  - SLU Link
       SLUPU    - SLU Physical Unit
       SLUSA    - Secondary Logical Unit Subarea
       TMOFFSET - Time Offset
       PSYHSA   - Host Subarea Address
       PSYIEXRT - Inbound Explicit Route Number
       PSYNODNM - Network Node Name
       PSYOEXRT - Outbound Explicit Route Number
       PSYSPUA  - SLU PU Address
       PSYTRPTY - Transmission Priority
       PSYVRRTE - Virtual Route Number

4. Following is  a brief discussion of NPM's techniques for
   measuring and recording various response time data.

   o Host response - the amount of time required to turn the
     transaction around in the host.  This time includes a
     small amount of VTAM time (NPM sits in the application
     subsystem interface) and all of the application (CICS,
     IMS, TSO) time.

   o Network response - to measure and record network
     response time, NPM requires that either (1) each
     transaction request a definite response (DR) from the
     terminal, or (2) the NPM DYNAMIC DR feature (available
     only with NPM) be used to force the transaction into DR
     mode.

   o Total response - the sum of host and network response
     times.

   There are, however, special considerations if this data is
   to be interpreted correctly.  NPM maintains buckets to
   save the data in the three categories above.  If definite
   response is neither requested nor forced for a particular
   transaction, there will be no network response time
   measure for that transaction.  If all transactions during
   a session interval are in definite response mode (network
   response is measured), total response, as recorded by NPM,
   is complete.

   However, when only some transactions within a session
   interval are in definite response mode, NPM computes the
   average network time (from the transactions that are in
   definite response mode) and adds this amount to the host
   time for each exception response mode transaction.  The
   result is an estimated total response time.  These
   estimates may or may not adequately reflect what the end
   user actually observed as total response time.

   This file maintains a data element, PSYPCNON (Percent of
   Total Events Without Network Response) that is intended to
   convey some confidence factor for the usefulness of the
   total response data.  For example, if 80% of the
   transactions within a session interval requested definite
   response, PSUPCNON will be 80%.

Retrieval Example

Chart response time for Primary Logical Unit (PLU) PRODCICS.
Break the chart bars into host and network response time.
Use yesterday's data.

   %LET BY = PLU HOUR ;
   %LET BREAK = HOUR ;

   PROC SORT DATA=&PSNTD..SNTPSY01 OUT=WKPSY;
     BY &BY;

   DATA WKPSY;
     SET WKPSY;
     %PSYSUM
   RUN;

   DATA WKPSY;
     SET WKPSY;
     IF PLU='PRODCICS';
     RESPONSE=PSYAVNTM; TYPERESP='NETWORK'; OUTPUT;
     RESPONSE=PSYAVHTM; TYPERESP='HOST   '; OUTPUT;

   PROC CHART DATA=WKPSY;
     VBAR HOUR / SUMVAR=RESPONSE
                 SUBGROUP=TYPERESP
                 DISCRETE;RUN;