cpvitu...@arkbluecross.com (Vitullo, Carmen P) writes: > I found this out some time ago working for Boeing, even though we were > one company, we still had to submit a budget each year for computing > services, this drove Boeing Helicopters to look at alternatives, > mostly the costs of CATIA and CADAM, on Big Iron, distributed was > initially cheaper, but in the long run many cost overruns due to poor > planning and a desire to 'just get off the mainframe at any cost', > that's what drove us to look at consolidation afterwards, then > migrations to / from other platforms....if it made sense :)
60s we started doing lots of stuff to leave system up 7x24 for online access. Part of the issue was that (especially initially), offshift online access was very light so there was little useage ... but in order to encourage offshit useage, the systems had to be available 7x24. Because of light useage there would be little recovery charges ... so lots of things were done to minimize offshift expenses. This was in the days when mainframes were leased and charges were based on system meter that ran whenever the processor and/or any channel was busy ... also everything had to be idle for at least 400milliseonds before it the system meter would stop (triva: long after mainframes had switched from leases to sales, MVS still had a time ever that went off every 400ms, making sure system meter would never stop). In any case, came up with special channel programs for terminal I/O ... that would let the channel "go idle" ... but would immediately startup when there was any characters. Also lots of support for "dark room" ... not requiring offshift operators. For big cloud megadatacenters, the price of systems has so dramatically dropped that they have hundreds of thousands of "blades" (each blade with more processing power than max. configured mainframe) supported by staff of 80-120 people per megadatacenter. Also with the dramatic cut in system cost, the major expense has increasingly become power & cooling. The big cloud datacenters have been on the leading edge of systems where power&cooling drop to near zero when idle ... but are effectively instant on for ondemand computing. while an undergraduate in the 60s, I was hired as fulltime boeing employee to help with the formation of boeing computer services ... consolidate all computing into independent business unit as part of better monetizing the investment (which would also have the freedom to sell computing services to non-boeing entities, a precursor to cloud computing). at the time, I thought renton data center was possibly the largest in the world with something like $200M-$300m (60s dollars) in ibm mainframe gear (for a period, there were 360/65s arriving faster than they could be installed, boxes were constantly being staged in the hallways outside the datacenter). there was disaster scenario where mt rainier warms up and the resulting mudslide takes out the renton datacenter. the analysis was that the cost of being w/o the renton datacenter for a week was more than the cost of the renton datacenter ... so there was a effort underway to repicate the renton datacenter up at the new 747 plant in everett. in any case, the politics with the different plant managers tended to dwarf any technical issues. -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
