[email protected] (Edward Gould) writes:
> Brian:
> One of there specific episodes we had was that the master console was
> the only one that was “talking”. As to other options you listed the
> auditors cut them off years ago and no use arguing with them (BTDT).
> As for HMC remote access again the auditors wouldn’t allow it no
> matter how much we argued (besides I sort of agree with them on the
> HMC issue).
> I won’t go into the old xmas party story I use regularly on here and what 
> could happen.
>
> Ed

For CP67 in the 60s to move to keeping the system available 7x24 a
number of things were done for online use. Part of it was reducing the
cost of operating off-shift ... dark room, no operator onsite, etc.

Initially offshift was very light, not justifying the cost of keeping
the system up ... but w/o 7x24 availability it wouldn't encourage
non-primetime online use.

Part of support was allowing "operator's console" to be other than the
1052-7 system console. Other part was auto fast failures with automatic
reboot and system up w/o manual intervention.

This was still in the days of when systems were leased with charges
based on the "system meter". The "system meter" ran whenver the
processor and/or any channels was busy. One trick was channel program
that would let channels go idle, but would instantly wake up for any
arriving characters. Another issue was that the "system meter" would
continue to run for 400milliseconds after all processing and channels
were completely idle. Lots of system work was to allow "system meter" to
stop (no rental/leased charges) when system idle, but immediately
activate when there was anything to do. Trivia: MVS had a system time
task that woke up every 400milliseconds long after business converted
from rent/lease to sales (guarenteed that system meter never stopped).

Another issue was that the science center had ported apl\360 to CMS for
CMS\APL ... and was letting other IBM locations use the system. The
business planners in Armonk loaded the most valuable of all corporate
data on the cambridge system to run business models. The security of the
system had to meet very high security standards (besides being available
7x24 and run dark room w/o human intervention) ... especially since
non-employees had online access to the system; students, staff,
professors, etc at universities in the boston/cambridge area.

past ibm cambridge science center posts
http://manana.garlic.com/~lynn/subtopic.html#545tech

There were some spin-offs from the science center that started using
CP67 to offer commerical online services (later migrating to VM370).
They were also expecially sensitive to both operating costs and
sensitive to security issues ... especially moving up the value stream
to the wallstreet financial community (where large competing financial
institutiosn were using the same systems).

past posts about commercial online (virtual machine based) services
http://manana.garlic.com/~lynn/submain.html#online

There are some similarities between what was being done for these
operations in the 60s and what the large cloud operators are doing now
in their megadatacenters (hundreds of thousand of systems with tens of
millions of processors per megadatacenter) for costs drop to near zero
when idle but immediately instant on when needed.

-- 
virtualization experience starting Jan1968, online at home since Mar1970

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to