Re: Re: slanted article "Inside System Storage -- by Tony Pearson"
Mike Schwab <[email protected]> Wed, 22 Sep 2010 22:51:19 -0500
| Newsgroups | gmane.comp.emulators.hercules390.advocacy |
|---|---|
| Message-ID | <[email protected]> |
On Wed, Sep 22, 2010 at 10:00 PM, Paul Raulerson <[email protected]> wrote: <deleted> > So let me turn the question around - what can we do with Hercules that would: > (1) Allow IBM to remain in control of their products and brands? OS/390 (after a certain point) and z/OS are able to detect that they are running on Hercules and if a certain configuration parameter is not set to a certain value it goes into a wait state shortly after IPL. z/PDT comes with a USB device that has a license key for a certain number of emulated processors. Hercules configuration file also provides the CPU serial number. Should be able to use existing z/OS code to obey the license capacity, perhaps with some modifications. The best x86 / x64 processors provide about 50 MIPS per emulated processor / x86 thread, so a 2 emulated processor on 1 core hyper-threaded processor (or higher) would verify that multiprocessing locks work while only providing 100 MIPS, about half the speed of a 2 processor z800 at a much lower price and electrical utilization. Incorporate the existing capacity utilization reports into the basic system so reports are automatically sent it, instead of the user having to explicitly run the report and FTP the results to IBM. Another function could be to ask about the computer's IBM license and IBM could program a licensed processor reduction or possibly even revocation of the license. > (2) Encourage or enable research or development into mainframe subjects? Use of PCs to work with mainframe software instead of z/800s would certainly be much cheaper, even just the electrical charges. > (3) Decrease internal costs of mainframe support and training? Micro Focus mainframe emulation was a great start. But this is hardware emulation running the actual z/Series software. > (4) Encourage widespread but controlled development of new mainframe software. They have been giving to universities and colleges traded in mainframes and dasd storage units and developed college courses to fill the need for retiring mainframe personel. These people tended tended to start working in their early 20s from 1960 - 1990 and the system programmers are all retiring after doing this for 40 years. This could be an option for the smaller universities who don't want to build a high power computer room. > (5) Encourage or enable the use of Linux on mainframes z/Linux is heavily run and developed on Hercules, at least 2,000 IBM employees. Enables them to do single step debugging without impacting anyone else. > (6) Enable embedded mainframe programs in very small devices? Somebody has compiled Hercules and installed it on an Linux / Atom / SD card operating system so that the 1 disk LAN file server became a Mainframe files server. Ivan Warren ran VM/370 5 pack on an IPaq PDA. > That's the kind of stuff we should be showing and advocating. That's the kind of stuff I believe will get us where we want to go. > > -Paul <deleted> The main advantage of Mainframes over Windows / Linux systems is the direct access to the needed record without having to open the entire file. Keeping the user logged in with a under 4K of disk space instead of MBs of ram dedicated to each user. Should Windows / Linux manage to build such applications their usage as mainframe replacements could explode. -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all?