Hi! As we know modern IO subsystems can handle single IO by 20-5 ms. I know there are many factors like caching, cylinders, etc etc, but lets ignore them in context of this discussion. Then single (Oracle) process can issued (1sec/20-5ms) 50-200 IOs/sec (sequential reads). Lets assume that process work with large tables, and main access path is thought indexes. Some time (depending of algorithm) process will spend for data processing, lets assume ~50%. The conclusion is single process can issue maximum 25-100 PIO per second, depending of IO subsystem. My question is, can it be rule of thumb, for peoples who going to tune some batch processes? What is the opinion of audience? Thanks in advance, Jurijs PS I assume, that process requires a lot of PIO-s by business. PIO is main bottleneck. PPS Going to particular system we can estimate average speed of IO system, STATSPACK reports, as well as from 10046 for particular process PPPS Next question will be, how practically we can test IO system before running production on it? PPPPS How to estimate TOTAL PIO count that IO system can handle? ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put 'unsubscribe' in the subject line. -- Archives are at //www.freelists.org/archives/oracle-l/ FAQ is at //www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------