[email protected] (David Andrews) writes: > Once upon a time I found it useful to condense a month's worth of RMF > data into a single graph showing the average CPU utilization over the > course of a day, plus-or-minus one standard deviation. That drop-bar > chart made it easy to visualize two-thirds of our daily workload at a > glance.
in the early days of commercial virtual machine online service bureaus (sort of the cloud computing of the 60s & 70s) ... there were reports showing the peaks & troughs of avg. daily online use ... and being able to extend use over the whole country ... allowing the peaks from the different timezones to offset the troughs in other timezones. the science center had started accumulating all its (virtual machine) system activity from the 60s ... and established it as standard process for normal operation. By the mid-70s, the science center not only several years of its own data ... but was also acquiring similar data from large number of internal datacenters. This was used for a lot of modeling and simulation work, along with workload & configuration profiling ... which eventually evolves into capacity planning. one of the science center's models (implemented in APL) ... was made available (starting in the mid-70s) on the (internal online virtual machine) HONE systems (providing world-wide sales & marketing support) as the "Performance Predictor". Sales people could collect customer workload & configuration profile and ask "what-if" questions (of the "Performance Predictor") about changes to workloads and/or configuration. misc. past posts mentioning science center http://www.garlic.com/~lynn/subtopic.html#545tech -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- 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
