[email protected] (Charles Mills) writes:
> LOL. Thanks. No, I really can't remember. Maybe too many illegal substances
> in the 1970's.
>
> I remember some of the places I bought time: Bayer in Emeryville, Central
> Bank Computer Bureau in Oakland, ..., but it was neither of those. (Man,
> those were different times! Can you imagine a drug company or a bank turning
> over their entire mainframe hands-on to some kid with long hair, with no one
> supervising what he did?)
>
> No harm in identifying the participants: they've probably all been promoted
> to management by now.
>
> (Someone wrote me off-line and asked if I meant "lights-out operations" or
> "lit-up operations.")

there was a lot of work on cp67 in the 60s to enable 7x24 online access.

one of the issues was that there was still significant shift costs
associated with enabling 7x24 access (both local and dial-up remote) ...
operators, lease/rental charge.

part of the issue was initially off-shift use was light ... so there was
little useage charges to recover fixed operational expenses.  in attempt
to minimize these there were lots of work done for operator-less
operations ... things like automatic ipl/startup ... w/o requiring
operator to be present ... allowing for lights-out operation (aka nobody
physically present in the room).

another big issue was that machines were rented/leased and charges were
based on the "cpu meter" that ran whenever the cpu and/or any channels
were operating ... including terminal i/o channel programs waiting for
dailup connections and/or people to type at their keyboard.  Solution
was to use the "prepare" CCW would allow terminal controller to
disconnect from the channel (allowing the cpu-meter to stop on otherwise
idle system) ... pending incoming connection and/or keystrokes (this is
slightly analogous to the disk "SET SECTOR" CCW that shows up later for
3330s&2305s with block-mux channels).

the result was not only was cambridge able to provide for cost effective
online 7x24 operation ... but also saw commercial service bureaus in the
60s starting to offer cp67 7x24 online computer access.

One of the things done by at least the mid-70s in the commercial service
bureaus ... although no shipped by IBM ... was non-disruptive,
loosely-coupled, process migration. they started offerring 7x24 online
operation around the world ... which by mid-70s also resulted in
requirements for 7x24 availability. A big problem then becomes scheduled
maintenance. Their solution was to have loosely-coupled, multiple system
operation ... and be able to take any system and/or compenent offline
... by non-disruptive (transparent) migration of all computing processes
from one system to another (non-ibm source-code system modifications).

recent discussion mentioning source code maintenace:
http://www.garlic.com/~lynn/2013e.html#66 Sequence Numbrs (was 32760?
http://www.garlic.com/~lynn/2013e.html#85 Sequence Numbrs (was 32760?
http://www.garlic.com/~lynn/2013e.html#88 Sequence Numbrs (was 32760?

trivia ... the cpu-meter required 400ms of idle (both cpu and all
channels) to coast to a stop. long after most hardware had shifted from
lease to purchase, MVS still had a fixed timer interval that would wake
the system every 400ms (making sure if the system was running at all,
the cpu-meter would never stop).

-- 
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