[email protected] (PINION, RICHARD W.) writes:
> I hacked my phone, installed Hercules, installed MVS 3.8, and 
> now my phone is controlled by MVS.
>
> But, I'm sure the Wheeler's would suggest I use VM/370 instead.

re:
http://www.garlic.com/~lynn/2017d.html#40 What are mainframes
http://www.garlic.com/~lynn/2017d.html#41 What are mainframes

in the mid-90s (having left IBM), I was brought into the largest airline
res system to look at the ten impossible things they couldn't do. I go
away and come back with all ten impossible things implemented (demo on
rs/6000 530). A big part was their existing implementation was still
based on technology trade-offs made during the 1960s. I could make
totally different trade-offs ... including making it run 100 times
faster. The processing for all passengers for all airlines in the world
could be handled by ten RS/6000 990s (a decade later, a cellphone had
the processing power and storage to it).

then the hang-wringing started. it turns out the part of 60s trade-offs
involved several hundred people manually prep'ing the data ... the redo
could use the full OAG directly w/o needing several hundred people
prep'ing the data for use.

however, the last product we did at IBM was (RS/6000) HA/CMP.
http://www.garlic.com/~lynn/subtopic.html#hacmp

We were working with national labs on cluster scaleup for filesystems
and scientific/technical and also with the (non-IBM) RDBMS vendors for
commercial. Old reference to JAN1992 meeting in (Oracle CEO) Ellison
conference room on (commercial) cluster scaleup
http://www.garlic.com/~lynn/95.html#13

within a few weeks, cluster scaleup was transferred, announced as the
IBM supercomputer (for scientific/technical *ONLY*) and we we were told
we couldn't work on anything with more than four processors. Possibly
part of the problem was that the (mainframe) DB2 group had been
complaining that if I was allowed to go ahead, it would be at least five
years ahead of them. Within a few months, we have left IBM.

17Feb1992 press, announced for "scientific and technical" *ONLY*
http://www.garlic.com/~lynn/2001n.html#6000clusters1 11May1992 press,
IBM "caught by *SURPRISE*" by national labs interest in cluster
supercomputers
http://www.garlic.com/~lynn/2001n.html#6000clusters2

in prior life, my wife was con'ed into going to POK to be responsible
for loosely-coupled (mainframe for "cluster") ... where she developed
peer-coupled shared data architecture ... some past posts
http://www.garlic.com/~lynn/submain.html#shareddata

however she didn't remain long because 1) poor uptake (except for IMS
hotstandby) until sysplex and parallel sysplex and 2) constant battles
with the communication group trying to force her into using SNA/VTAM for
loosely coupled operation (there would be periodic temporary truce where
they said she could use anything within the walls of the datacenter, but
they had corporate strategic ownership of everything that crossed the
datacenter walls, but then communication group would break the truce and
start again).

as an aside, early 1979, I was con'ed into doing benchmarks on
engineering 4341 for LLNL that was looking at getting seventy for
compute farm ... leading edge of the coming cluster supercomputing
(cluster supercomputing interest by national labs from more than decade
earlier).

One of my hobbies (70s through mid-80s) was enhanced operating systems
(develop, ship, support) for internal datacenters ...  one of the long
time customers was world-wide online sales&marketing support HONE
system. In the mid-70s, the US HONE datacenters were consolidated in
Palo Alto (trivia: when facebook first moved into silicon valley, it was
into a new bldg built next door to the former HONE datacenter). In 1979,
HONE had largest single-system-image, loosely-coupled mainframe
operation in the world with load-balancing and fall-over. In the early
80s, it was replicated in Dallas and then in Boulder with fall-over
between datacenters.
http://www.garlic.com/~lynn/subtopic.html#hone

some airlines might have ACP 8-way loosely-coupled (3830 4channel switch
with 3330 string switch for up to 8 channels) but didn't get
tightly-coupled support until many years later.  US HONE had 8-way
loosely coupled and each processor complex had 2nd "attached-processor"
(for 16 processors total). Whole device reserve/release was performance
killer. ACP did have the ACP 3830 "lock" RPQ ... supporting fine-grain
logical locking ... but didn't work across 3830 controllers with 3330
string switch (limited to 4channel configurations). HONE instead used a
channel program sequence that was the logical equivalent to the
compare&swap instruction. past posts mentioning SMP and/or compare&swap
http://www.garlic.com/~lynn/subtopic.html#smp

"From Annals Of Release No Software Before Its time" 2009 post about IBM
press regarding RS/6000 RDBMS cluster scaleup (nearly 20yrs later) and
zVM loosely-coupled operation (30 yrs later)
http://www.garlic.com/~lynn/2009p.html#43
http://www.garlic.com/~lynn/2009p.html#46

-- 
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