Chapter 5. Database performance 121
򐂰 Line 12: NUMBER OF TIMES CTRL INTERVAL REQUESTED ALREADY IN POOL
This count indicates the number of times a request for a CI was satisfied from the subpool
and therefore a read operation was not required. It is used to calculate a hit ratio for the
subpool (discussed under 5.20.4, “VSAM statistics” on page 121 at the end of this list).
򐂰 Line 13: NUMBER OF CTRL INTERVALS READ FROM EXTERNAL STORAGE
The count reports the number of requests that could not be satisfied from the contents of
the subpool and had to be read in from a database data set on DASD.
Use tuning tactics to attempt to eliminate reads or, if they cannot be avoided, to perform
the reads in as efficient a manner as possible. This line item is used to calculate a hit ratio
for the subpool (discussed in 5.20.4, “VSAM statistics” on page 121).
򐂰 Line 14: NUMBER OF VSAM WRITES INITIATED BY IMS/ESA
The count reports the number of CIs written to database data sets as a result of
application program sync points (including CHKP and SYNC calls) and background write.
򐂰 Line 15: NUMBER OF VSAM WRITES TO MAKE SPACE IN THE POOL
The count reflects the number of CIs written to database data sets as a result of altered
buffer steals.
򐂰 Lines 16 and 17: NUMBER OF VSAM READS FROM HIPERSPACE BUFFERS and
NUMBER OF VSAM WRITES TO HIPERSPACE BUFFERS
The reads (Line 16) reflect a request of the buffer handler that was satisfied from
Hiperspace. The writes (Line 17) reflect the movement of CIs from the virtual storage
subpool buffers to Hiperspace. Every read from DASD or Hiperspace into the virtual
storage subpool results in a write to Hiperspace.
The ratio of writes and reads to and from Hiperspace is an indication of its effective use.
Reasons for using Hiperspace rather than virtual storage buffers include these:
A VSAM subpool in virtual storage is limited to 32 KB buffers. But there is no practical
limit to the number of Hiperspace buffers that can be used.
From a buffer handler point of view, SCHBFR (NUMBER OF VSAM SCHBFR CALLS
ISSUED, Line 11) requests consume CPU resource; the more buffers in a subpool, the
longer the path length. SCHBFR requests do not search Hiperspace buffers.
Therefore, if SCHBFR requests are perceived to be a problem, placing most of the
buffers in Hiperspace reduces the path length required to satisfy the requests.
For HIDAM, a better solution is to provide enough free space so that the most desirable
and second-most desirable CIs have enough space to satisfy requests and then to
schedule timely database reorganizations to buy back that free space when required.
VSAM reads from Hiperspace buffers are used to calculate a hit ratio for the subpool
(discussed in 5.20.4, “VSAM statistics” on page 121).
򐂰 Lines 18 and 19: NUMBER OF FAILED VSAM READS FROM HIPERSPACE BUFFERS
and NUMBER OF FAILED VSAM WRITES TO HIPERSPACE BUFFERS
Counts in either of these two fields indicate that storage is overextended. Hiperspace, as
implemented by VSAM, does not guarantee that Hiperspace buffers are not lost.
5.20.4 VSAM statistics
Several statistics are of use when analyzing a VSAM database subpool. The following are
some of the more common statistics used:
򐂰 Hit Ratio
A subpool hit ratio is calculated as follows:

Get IMS Performance and Tuning Guide now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.