> On Jan 13, 2018, at 12:18 PM, Ron hawkins <[email protected]> wrote: > > Ed, > > The way I remember it is that not long after the first 3390 started to ship, > IBM "promised" that they would not change track geometry again.
Maybe we are talking about the same thing but from different POV’s ? I can recall at least 3 Guide presentations and several IBM presentations that basically indicated that *IF* you use the SDB *and* the new way of allocation of space in Megabytes and or records, the geometry of the device would be essentially invisible. The three IBM/GUIDE presentations were from WSC and some place else inside IBM. This goes back what 25 years or so. Can I prove it, no. But if you do use the new way, indeed track geometry doesn’t effect anything. Having said that the wild card that I don’t think even IBM thought about was the 512 byte (or some combination there of) FB architecture. I am guessing here, but it would seem logical to me that using the new method of allocation, IBM could and probably will make it transparent to every one. 10-15 years before SMS, either IBM came out and gathered data (SMF and disk maps etc) from us. They wouldn’t say much, my vague memory says that our IBM SE helped them out gathering the data. Our IBM SE also did an “orange” manual (its been ages and I don’t remember the color) study on tape. I think that is what got him noticed from the WSC. When he left us he moved to Gaithersburg MD and then out to San Jose. The only hint our SE gave us was that something new was coming down in the pipeline. His name was Jim Garner and he was an excellent IBM asset. I kept running into him at Guide and we always went out to dinner during Guide. Jim never talked about specifics but it was fun to hear about items being talked about in the future. On the other end, I had a very good friend in IBM that used to work in the White House computer system (I think Profs but again I could be wrong). In *think* he was in the WH when Watergate happened and he spent a lot of time trying to recover “email”. My friend was telling me that because of the “bunkers” IBM could not use standard channel cable and had to use Fiber optics. I remember him talking about it as error handling was different and EREP had to be updated to allow for it. About 20 years afterwords it showed up available for use. I never tried to pry but let him talk. He is now long time retired up in Maine. I still hear from him once or twice a year. > > The proof of the pudding was that the real 3390-9 kept the same track > geometry, using three logical tracks on a physical track when there was an > opportunity to go for something more like 64KiB. > > The same opportunity was there when the ESS shipped, but they have stuck with > their promise. I would not call SDB an SMS thing. It has helped with changes > in optimal blocking for EF data sets, but I don't recall it being protection > against changing form factors. I don't know that DFSMSdfp was ever going to > deliver that while the track size is in the Base config. > > Ron > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On > Behalf Of Edward Gould > Sent: Thursday, January 11, 2018 7:21 AM > To: [email protected] > Subject: Re: [IBM-MAIN] Number of Cylinders per Volume > >> On Jan 11, 2018, at 6:44 AM, R.S. <[email protected]> wrote: >> >> Ed: >> I am sure I had read that. Of course such promise can be broken in the >> future (why not?), but whole idea of so called mod-27, mod-54 and finally >> EAV was to extend volume size with no change to geometry. >> >> (speculation mode) >> IMHO the most likely change in the future we can expect is CKD - > FBA >> migration. Unless FBA storage world will change current FBA model to >> something, let's say "more flash lilely". For example FBA world changed >> (actually came back *) the sector size which was 512B net for ages and now >> it's commonly extended to 4096kB. That introduces some incompatibility >> problems. >> (*) bigger sector sizes were used in magneto-optical removable media drives. >> With big pain for some Unix systems and not only. >> >> -- >> Radoslaw Skorupka >> Lodz, Poland > > Radoslaw: > > I was a project manager at Guide in the 90’s for storage. I vaguely remember > being invited to San Jose (as part of SHARE+side story at bottom). IBM > supposedly laid out their 10 year plan for DASD. Here is where it gets a > little fuzzy. I believe the “future” was to be SMS PERIOD. IBM during one of > the small sessions said that by being completely SMS you would no longer care > about device geometry as long as you used the IBM SDB and the appropriate > space allocations. I was a "not in my lifetime person after the session". I > talked with the other project managers and they had reservations as well but > after years and years of changing track size and block size issues and > getting gray hair along the line I think we were happy to finally rid > ourselves of one more PITA. At the time (IIRC) not all of the new SMS > features were available. We were told to hold on it was being worked on. > > To complete the story, I am a semi believer never made it to 100 percent > believer. > > Now the side story: One (or two?) project managers were employed by OEM > vendors. IBM put a restriction of attendance that if you were a OEM vendor > you couldn’t attend. Its been so long that I do not remember if IBM backed > down or the people were disinvited or what. > > Ed > > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, send email to > [email protected] with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
