I have been involved in a number of offline discussions as well as multiple IBM 
PMRs and problem records with our DASD hardware vendor concerning EOS and 
although this is potentially a most useful function, the overhead involved is 
not zero--there will always be *some* amount of overhead involved in re-writing 
every bit of a data set, no matter how much improvement IBM has made to the 
function at the OS level (such as the 95+% reduction in EXCP's since the 
initial implementation of EOS).  During our testing, we find that it takes an 
additional 1 hour of processing time for every 500G of data in order to use 
EOS.  If you have a tight batch window or any kind of time-sensitive 
processing, you should make sure that you can afford to use EOS before simply 
activating it for every data set in your enterprise.  Since this is a 
security-related function, I will not comment publicly if/when/where we do or 
do not use it, but I will simply recommend that you perform a benchmark 
evaluation of the overhead involved on your hardware before activating EOS.

Larre Shiller
US Social Security Administration

"The opinions expressed in this message are mine personally and do not 
necessarily reflect the opinion of the US Social Security Administration and/or 
the US Government."

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to