VSAM KSDS datasets were a clear win as a replacement for Indexed
Sequental (ISAM) datasets when adding large numbers of keyed records. I
saw cases where a KSDS implementation literally ran two orders of
magnitude faster than ISAM and also took less DASD space, because ISAM
required that unblocked overflow records be serially searched when
there was no remaining space in a block to insert a record with a given key.
In theory an ESDS dataset could be used to replace a sequential dataset
and BSAM/QSAM, but of course the application interfaces were
considerably different, and VSAM constraints on block size meant you
could take substantial hits on track efficiency and performance for
certain logical record sizes.
An RRDS could be used as a functional replacement for direct access
files, but again the restriction on block sizes caused compatibility
issues, and tuning RRDS dataset access to get performance comparable to
a well-tuned BDAM application was difficult to impossible.
You could probably have designed a functional replacement for PDS
datasets with either a KSDS or RRDS organization, or a combination of
KSDS and ESDS, but it wouldn't have been practical. Many decades later
PDSEs are almost a tranparent replacement for a PDS, but there are still
some things possible with a PDS that are not supported for a PDSE.
If VSAM's goal was to replace all other file organizations, it failed.
The only old dataset organization to be made totally obsolete by VSAM
was ISAM.
JC Ewing
On 5/24/24 10:02, Paul Gilmartin wrote:
On Thu, 23 May 2024 22:24:06 -0500, Mike Schwab wrote:
VSAM came from the Future Systems development as a complete
replacement, Lynn Wheeler has posts about that.
It was cut back to be an addition to MVS, then combined with CVOL
catalogs to ICF.
"complete replacement" of what, specifically? I have heard the
assertion that VSAM was intended to replace all other access
methods: QSAM, BSAM, BPAM, ...
...
--
Joel C. Ewing
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
