DASDBILL2 wrote:
It is my cold and fuzzy recollection that 6144 was in vogue for a few years
when that value was deemed to be VERY good on a 3340 DASD and not too shabby
[1] for most other flavors.
Bill Fairchild
[1] Or, more precisely, "okay" or "swell".
<snip>
If you have your old, yellowing DASD reference cards (everyone saves
*those*, right?), then you can see that ~6K was a decent submultiple of
the track lengths of the day for most devices. As a compromise it had
its merits, since if I recall correctly (I am far too lazy to look it
up, sorry), it had track utilization in the high 85% or better range on
all the devices of the day. This led to common block sizes like 6144
for load libraries and 6160 for FB80 data sets.
When we did DASD conversions every few years, this made eminently good
sense. Copying from one device type to another without reblocking did
not completely destroy space utilization and cause data sets to take
hugely unexpected gobs of disk space. In other words, from a usability
standpoint they were very good choices indeed.
However, these block sizes were *never* optimum on any device, as Siebo
Friesenborg pointed out in some detail a long time ago in various
papers, presentations, books, and such. (Siebo and I corresponded in
the late 90's about the special case represented by load libraries.)
--
John Eells
z/OS Technical Marketing
IBM Poughkeepsie
[email protected]
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
