On Mon, 10 Apr 2006 00:00:00 GMT, Ted MacNEIL wrote: >This may sound like a CICS question, but it is about the cost of doing I/O. > >We have a CICS application that is doing upwards of 2000 I/O's per second >to a single VSAM file. >The entire CICS region is using 4.5% of a z/990 2084-309. >The CICS buffer hit ratio is around 60%, after we doubled the number of >buffers (can't do that too many times). > >The cache hit ratio is over 98%. > >We have run out of all tuning options, save one. >Replace or remove the application.
Ted, How large is the VSAM file? Can it fit into a CICS Maintained Data Table? We found that, by moving a highly-active VSAM file into a CMDT that we were able to perform over 4.5 BILLION I/O operations to the file each day! I then wrote an I/O trace that proved that the application did not NEED to access the file as much as it was -- by a factor of over 100 -- and reduced the file's I/O rate to much less than our busiest (other) files. We learned two things: CMDTs are remarkably robust and our application people weren't paying attention to important details. (Both of those were intuitive but its nice to have proof every once in a while.) The I/O trace was run from XFCFROUT where we could look at the record keys - - if you read and then re-read the same key from the same CICS task (and do that again and again)... you ARE the weakest link! Also, if you browse through records you have no need to look at in order to find those records you intended to browse... you get the drill. We got a boatload of CPU cycles back in the process, of course. -- Tom Schmidt Madison, WI (Lest anyone think that couldn't happen at their site -- wanna bet?) ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
