In our environment Shark 800 w/ficon it's 2 ms or less into the hundreds of mic's with cache hits in the 90 - 100% range. When we supported RVA T82, w/escon w/3390 emulation I seem to remember 4 - 7 ms with same cache hit ratio. I also found some old doc from out site that showed STK 8890 with 48 mg cache with read hit ratios of 80 - 100% getting between 5 - 10 ms. And if memory servers me, I'm still looking for the doc, on IBM 3390 w/3990 w/o cache I would say anywhere from mid teens to high 20's ms. response time but of course results will vary with pathing, workloads and the like.
David Day <[EMAIL PROTECTED]> Sent by: IBM Mainframe Discussion List <[email protected]> 08/23/2006 03:27 PM Please respond to IBM Mainframe Discussion List <[email protected]> To [email protected] cc Subject DASD Response Time I'm developing some code on a Flex-es machine and would like to try to see how the code will work in a real environment. I don't know if it is Flex-es or Linux that is caching the I/O, but the response time the application sees is extremely good. The Flex system has the ability to delay the I/O that it delivers to MVS. I would like to set these delay values to something that more resembles actual I/O response times. If someone can give me some values to plug in, I'd appreciate it. 3390's, with and without a caching controller. I know the values vary with cached controllers and hit ratios. I'm interested in the response time RMF reports when the data is found in the cache. Thanks. --Dave Day ---------------------------------------------------------------------- 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 ---------------------------------------------------------------------- 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
