IBM Mainframe Forum

I have an online file that is used in a batch job and I wonder if there is a way to speed up processing by adding different buffering. Here is what we have coded for the JCL now:

//CT3NAARC DD DSN=XXX.XXXXXXX.XXXNAARC.ONLIV,DISP=SHR,
// AMP=('BUFNI=4,BUFND=50')

Can you use the BLSR parameters to buffer these datasets? Is there another way to buffer these? Hiperspaces? etc...

Hello,

If the online is "pounding away" on this file and some batch process is permitted to run at the same time, both will be impacted.

Suggest you implement some kind of scheduling that will ensure the batch job is not in competition with the online for the file.

If you believe the batch job needs improvement when it runs unopposed, additional buffers may help. You would also need to consider what kind of "things" are being done in the batch job.

Based on the existing buffering, your batch job is using sequential access. BLSR is really bad with sequential access, so using it could actually make your batch job performance worse. As Dick indicated, your best bets are to add more buffers and to move the batch job schedule so there's no conflict with online.

BLSR goes well with VSAM I/O.