stars...@mindspring.com (Lizette Koehler) writes: > https://www.ibm.com/blogs/cloud-computing/2014/03/a-brief-history-of-cloud-compu > ting-3/ > > After some time, around 1970, the concept of virtual machines (VMs) > was created.
mid-60s, some of the CTSS people went to 5th flr to do MULTICS ... others went to the ibm science center on the 4th flr and did cp40/cms (after having done the hardware modifications to add virtual memory to 360/40). cp40/cms morphs into cp67/cms when 360/67 that came standard with virtual memory becomes available in 1967. transition to online available 7x24 was an issue since machine useage offshift from home was quite variable ... and ibm mainframe even sitting idle (but available) was quite expensive. lots of work was done on cp67 to support darkroom, unattended operation ... to minimize offshift costs when (especially initally) there was little useage (but in order to promote offshift useage, system had to be up 7x24). this was also in the days when systems were rented ... and monthly charges was based on the "system meter" that ran whenever the processor and/or (any) channel was executing (processor and all channel activity had to be idle for at least 400ms before system meter stopped). cp67 did some special programming so that channel would go idle ... but instantly wake up whenever there was any arriving characters ... further reducing offshift costs when idle. Trivia: long after the shift from rent/leased to sales ... MVS still had a timer task that woke up every 400ms (making sure that system meter never stopped). The science center offered online service both internal and to students/staff/professors at local univ. in cambridge area. One of the issues was highest security since some of the internal users were Armonk business planners which had loaded the most valuable and sensitive corporate data on the system ... and the system was also being used by non-employees from local universities. CP67/CMS was also being used by various gov. agencies. I was undergraduate at the time but doing extenive CP67/CMS changes ... and would even periodically get requests from IBM for enhancements. I didn't know about it at the time, but some of the (security related) requests from IBM may have actually originated from gov. agencies (although I didn't find out about them until much later). gone 404, but still lives free at the wayback machine: http://web.archive.org/web/20090117083033/http://www.nsa.gov/research/selinux/list-archive/0409/8362.shtml even before I graduated, Boeing hired me full time to help with the formation of Boeing Computer Serivces ... consolidate all of dataprocessing in an independent business unit to better monetize the investment (including offering services to non-Boeing entities). Boeing Renton had somewhere around $300M (late 60s $$s) in IBM mainframes ... 360/65s were arriving faster than they could be installed ... and were in the processor of replicating Renton datacenter at Paine Field (for disaster survival). late 70s/early 80s customers bought large number of 4300s. datacenter 4300 clusters had more aggregate processing & i/o than high-end mainframes at significant lower cost, smaller footprint, lower power useage and environmental requirements. large customers also had orders of 4300 in hundreds at a time for placing out in departmental areas (sort of the leading edge of the distributed computing tsunami). In 1979, I was asked to do 4341 benchmarks for LLNL that was looking at getting 70 4341s for compute farm ... a leading edge of the coming cluster supercomputers. modern large cloud operators will have several megadatacenters, each one with hundreds of thousands of systems (and millions of processors) staffed by 80-100 people ... to meet elastic, on-demand computing. They claim that they assemble their own systems for 1/3rd the cost of systems from brand name vendors (and server chips for cloud megadatacenters exceed number going to brand name server vendors) ... enabling provisioning for enormous elastic ondemand ... system costs have been so radically reduced that power&cooling have increasingly became major cost. They've worked extensively with chip makers so that they can have systems where power/cooling drops to zero when idle ... but are "instant on" as needed for on-demand use (going way beyond what was done for cp67/cms in the 60s). -- 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