If there is no queuing on the channel/cu/device), then who cares if it is 100% busy?
You just need to be aware that at 100% busy, you can't (ever) go any faster. If SLA's are being missed, then you either need more of the same hardware, or an upgrade to faster hardware ($$$$). Many of the old "ROTs" (Rules of Thumb) had a theoretical basis in the hardware designs of the time. The hardware designs have advanced, and changed the theoretical basis, but we are still using the "old ROT's. Base your decisions on what (if anything) is being delayed in the hardware path, not on the ROT!!! The above discussion accounts for the "front end" of the VTS. The "back end" of the VTS is also an issue. I cannot speak to the specifics of a 3494 since I don't have one. If the time to retrieve a virtual volume is increased, this indicates either a lack of back end capacity, or VTS cache thrashing. <snip> >I have searched manuals, the web and IBMLink and cannot find a number for what is considered a limit for Escon channel usage. >Does anyone know what that number would be? If so, is it documented anywhere? When I used to do Tape Capacity Planning, the limit was 100%. I have not seen a similar analysis for ESCON (or FICON), but do have one question. Is there a throughput bottleneck? If not, then don't worry about it. </snip> ---------------------------------------------------------------------- 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
