alan_altm...@us.ibm.com (Alan Altmark) writes: > I'm not sure what you're saying. MVS, VM, and VSE code bases *all* > precede the invention of channel-attached FBA. They weren't > engineered for use by MVS (e.g. originally no RESERVE/RELEASE), but it > didn't matter since MVS wasn't engineered to accept device geometries > that weren't based on (CYL, TRK, REC) addressing and allocation units.
re: http://www.garlic.com/~lynn/2013.html#30 Searching for storage (DASD) alternatives CP/CMS used ckd search paradigm as if it was fixed-block ... so when real FBA came along, it was trivial to remap to fixed-block. Note that a lot of CP/CMS had heavy influence from MIT CTSS/7094 ... which predated 360 CKD. ctss reference http://en.wikipedia.org/wiki/Compatible_Time-Sharing_System aka some number of the CTSS people went to the 5th flr to do Multics and others went to the IBM science center on the 4th flr and did virtual machines, internal network, bunch of other stuff. misc. past posts http://www.garlic.com/~lynn/subtopic.html#545tech OS/360 made heavy use of CKD multi-track search especially for vtoc and pds directories. I've frequently pontificated it was mid-60s trade-off between real-storage to maintain the information and channel/controller/device resource to perform the search outboard .... and the trade-off had inverted by the mid-70s; I would even get called into OS/VS2 multi-system accounts that were experiencing serious throughput problems because of the heavy used of multi-track search ... recent post about getting called into large national retailer ... after all the usual POK experts had been tried http://www.garlic.com/~lynn/2013.html#25 I've also periodically mentioned that I was told that even if I provided MVS with integrated and fully-tested FBA support ... that I still needed to show a $26M business case to cover education, documentation, training ... basically several hundred million dollars in incremental FBA disk sales ... and specifically could not use total total lifecycle savings ... and by-the-way ... customers were buying disks as fast as they could be produced ... so any FBA support would result in just changing from CKD sales to FBA sales ... not incremental new sales. This is despite the fact that all DASD was heading in the direction of FBA ... furthermore real CKD hasn't been manufactured in decades ... and just getting initial ECKD hardware working (to pickup a little of FBA benefit) cost on par with what they quoted me for MVS FBA support. misc. past posts mentioning CKD, FBA, multi-track search, etc http://www.garlic.com/~lynn/submain.html#dasd Of course, I've had somewhat similar encounter in the early 90s over fiber-channel support ... I had been asked in the 80s to help LLNL standardize some serial stuff that they had .... that eventually morphs into FCS in the early 90s. Then some POK channel engineers get involved and layered some heavy-weight stuff on-top of FCS that eventually becomes FICON ... and enormously reduces throughput ... compared to the native/underlying FCS throughput. recent post discussing fcs, ficon, & z196 max i/o benchmark http://www.garlic.com/~lynn/2013.html#10 From build to buy: American Airlines changes modernization course midflight -- 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
