From: lynn@garlic.com (Lynn Wheeler) Date: 22 Jan, 2011 Subject: America's Defense Meltdown Blog: Boyd Strategyand article I've referenced before ... beltway bandits and gov. contracts discovering that it is more profitable to succeed than to fail Success of Failure
aka the gov. considers it so important ... that it will keep the money flowing ... and a stream of failures just means more money (than simple success upfront). The reference to "wall street venality dwarfs pentagon" equivalent ... is too big to fail ... put all the bets on very high risk with enormous odds/payoffs ... but in case of big losses, the gov. is there with its never ending flow of money.
re:
https://www.garlic.com/~lynn/2008s.html#20 Five great technological revolutions
https://www.garlic.com/~lynn/2009e.html#53 Are the "brightest minds in finance" finally onto something?
https://www.garlic.com/~lynn/2009o.html#25 Opinions on the 'Unix Haters' Handbook'
https://www.garlic.com/~lynn/2009o.html#41 U.S. house decommissions its last mainframe, saves $730,000
https://www.garlic.com/~lynn/2010b.html#19 STEM crisis
https://www.garlic.com/~lynn/2010p.html#78 TCM's Moguls documentary series
https://www.garlic.com/~lynn/2010q.html#5 Off-topic? When governments ask computers for an answer
https://www.garlic.com/~lynn/2010q.html#69 No command, and control
https://www.garlic.com/~lynn/2011.html#55 America's Defense Meltdown
https://www.garlic.com/~lynn/2011.html#75 America's Defense Meltdown
https://www.garlic.com/~lynn/2011.html#93 America's Defense Meltdown
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: History of copy on write Newsgroups: comp.arch Date: Sat, 22 Jan 2011 14:58:46 -0500nmm1 writes:
360 ... all the storage operands would be pre-checked ... both starting and ending (for at least store & fetch & 360/67 for page fault) ... before starting the instruction ... and abort the instruction w/o doing anything. One of the byte-by-byte features was to propagate value thru a field using overlapping operands, place a zero in 1st byte of source operand, and then do a move with the start of the target operand at +1 (the 2nd byte of the source is the same as first byte of the target, the zero from the 2nd byte of the source isn't there until it had been moved there). some more recent machines will attempt to optimize with larger chunks if not overlapping operands.
370 ... introduced the "long" instructions ... which were defined to be incrementally executed a byte at a time, interruptable, and restartable. some of the early implementations would precheck the long instructions operand addresses ... and abort w/o executing (instead of incrementally executing until the problem address happened) ... which sometimes could be unpredictable results.
somewhere recently somebody "APAR'ed" the TR/TRT instructions ... which got it changed. The TR/TRT takes each byte of the first operand, indexes displacement into the 2nd operand and either replaces the 1st operand with the contents of the indexed byte (or stops because of condition). The default was assume that the 2nd operand was always a 256byte table and precheck both the starting and ending storage locations (for things like fetch protect). The APAR was that the 1st operand/source might contain a limited set of values and programmer takes advantage of the fact to build a table much less than 256 bytes (always checking 2nd operand +256 could give erroneous results).
newer implementations now check to see if the (table) 2nd operand +256 bytes crosses 4k boundary (page fault, fetch protect, etc) ... if it doesn't, the instruction is executed as before. If +256 crosses 4k boundary, the instruction is pre-executed to see if any (source) byte value results in address crossing a 4k boundary. New "performance" recommendation is to never place starting address of 2nd operand/table within 256 bytes of 4k boundary.
one of the digressions with regard to segment versus page protect. In the original (370 virtual memory) segment protect design ... there was a single page table (per virtual segment) ... with the possibility that some virtual address spaces (sharing the segment) had r/o, store protect and other virtual address spaces (sharing the same segment) had r/w access.
with segment protect, all virtual address spaces could utilize the same physical page table ... with the protection specification back in each virtual address space specific segment table entry (i.e. extra bit in pointer to page table). with page protect ... the protect indicator is in the (shared) page table entry (rather than in the non-shared segment table entry) ... resulting in all virtual address spaces sharing that (segment/) page table ... will have the same protection.
semi-related recent discussion about virtual memory protection
mechanisms versus (360) storage key protection mechanims:
https://www.garlic.com/~lynn/2011.html#44 CKD DASD
https://www.garlic.com/~lynn/2011.html#74 shared code, was Speed of Old Hard Disks - adcons
https://www.garlic.com/~lynn/2011.html#79 Speed of Old Hard Disks - adcons
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Rare Apple I computer sells for $216,000 in London Newsgroups: alt.folklore.computers, aus.electronics, aus.computers Date: Sat, 22 Jan 2011 16:57:42 -0500Ahem A Rivet's Shot <steveo@eircom.net> writes:
https://www.garlic.com/~lynn/96.html#4a John Hartmann's Birthday Party
https://www.garlic.com/~lynn/2000g.html#40 No more innovation? Get serious
https://www.garlic.com/~lynn/2002f.html#19 When will IBM buy Sun?
https://www.garlic.com/~lynn/2002g.html#79 Coulda, Woulda, Shoudda moments?
https://www.garlic.com/~lynn/2002o.html#33 Over-the-shoulder effect
https://www.garlic.com/~lynn/2003e.html#26 MP cost effectiveness
https://www.garlic.com/~lynn/2003f.html#13 Alpha performance, why?
https://www.garlic.com/~lynn/2004f.html#16 Infiniband - practicalities for small clusters
https://www.garlic.com/~lynn/2005p.html#23 What ever happened to Tandem and NonStop OS ?
https://www.garlic.com/~lynn/2005q.html#9 What ever happened to Tandem and NonStop OS ?
https://www.garlic.com/~lynn/2005q.html#36 Intel strikes back with a parallel x86 design
https://www.garlic.com/~lynn/2006l.html#39 Token-ring vs Ethernet - 10 years later
https://www.garlic.com/~lynn/2006y.html#31 "The Elements of Programming Style"
https://www.garlic.com/~lynn/2007f.html#17 Is computer history taught now?
https://www.garlic.com/~lynn/2007j.html#49 How difficult would it be for a SYSPROG ?
https://www.garlic.com/~lynn/2007n.html#21 The Development of the Vital IBM PC in Spite of the Corporate Culture of IBM
https://www.garlic.com/~lynn/2007n.html#86 The Unexpected Fact about the First Computer Programmer
https://www.garlic.com/~lynn/2007p.html#35 Newsweek article--baby boomers and computers
https://www.garlic.com/~lynn/2007v.html#53 folklore indeed
https://www.garlic.com/~lynn/2008e.html#8 MAINFRAME Training with IBM Certification and JOB GUARANTEE
https://www.garlic.com/~lynn/2008p.html#36 Making tea
https://www.garlic.com/~lynn/2008r.html#68 New machine code
https://www.garlic.com/~lynn/2009e.html#58 When did "client server" become part of the language?
https://www.garlic.com/~lynn/2010.html#15 Happy DEC-10 Day
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Rare Apple I computer sells for $216,000 in London Newsgroups: alt.folklore.computers, aus.electronics, aus.computers Date: Sat, 22 Jan 2011 21:24:01 -0500"SG1" <lostitall@the.races> writes:
https://en.wikipedia.org/wiki/Gary_Kildall
gone 404 ... but lives on the wayback machine
https://web.archive.org/web/20071011100440/http://www.khet.net/gmc/docs/museum/en_cpmName.html
kildall using cp67/cms at npg school (wiki mentions he fulfilled draft obligation by teaching at npg)
melinda's virtual machine history going back to science center, cp40 &
cp67.
https://www.leeandmelindavarian.com/Melinda#VMHist
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Mainframe upgrade done with wire cutters? Newsgroups: alt.folklore.computers Date: Sun, 23 Jan 2011 10:01:04 -0500hancock4 writes:
besides software, it wasn't unusual for there to be a "team" of SEs
("system engineers") assigned at bigger customers; nearly always onsite
at the customer to provide customer with whatever assistance was needed
for using the computer. With unbundling, these SEs services also became
"charged for". One of the big issues was lots of SE education had been
as sort of journeyman/trainee part of these SE "teams" onsite at
customer installations. With unbundling, nobody was able to figure out
what to do with "trainee" SEs (since if they were doing anything at
customer site, it had to be a billable line item). The "HONE" systems
were several internal (virtual machine) CP67 datacenters, initially for
providing "Hands-On" online access for branch office SEs ... being able
to practice their operating system skills. misc. past posts mentioning
HONE:
https://www.garlic.com/~lynn/subtopic.html#hone
one of the big (leased/metering) issues for offering online, 24x7
timesharing service was programming tricks to minimize meter running
when there was no activity (default was that meter would still run if
system was active/available ... even if it wasn't executing). early on,
off-shift use was extremely sporadic ... but it wasn't likely to
incerase ... unless the system was available, on-demand, 7x24 ... but
recoverable charges for the light, sporadic use wasn't sufficient to
recover hardware billing charges (having the meter only run when there
was actual use went a long way to being able to deploy 7x24 online
offering). some past posts about early 7x24 online commercial
timesharing services
https://www.garlic.com/~lynn/submain.html#timeshare
the company was able to make the case with the gov. that "kernel" software was still free (necessary for the hardware to operate).
the company then had the failed Future System effort ... some past
posts
https://www.garlic.com/~lynn/submain.html#futuresys
During Future System (which was going to completely replace 360/370 and be radically different), 370 efforts that were considered possibly competitive were killed off. Then with the demise of FS, there was mad rush to get items back into the 370 hardware & software pipelines. The lack of 370 products is also considered reason that clone processors were able to get market foothold. With getting new 370 items back into the product pipelines and the clone processor competition, there was decision made to transition to charging for kernel software.
I had been doing 360/370 stuff all during the FS period (and making
uncomplimentary comments about the FS activity). some old email (one of
my hobbies was providing packaged production enhanced operating systems
for internal datacenters):
https://www.garlic.com/~lynn/2006v.html#email731212
https://www.garlic.com/~lynn/2006w.html#email750102
https://www.garlic.com/~lynn/2006w.html#email750430
With the demise of FS and mad rush to get out 370 products, various
pieces that I had been doing were selected to ship to customers. Some of
the items selected were to be packaged as (kernel add-on) "Resource
Manager" product. My "Resource Manager" then got selected to be guinea
pig for starting to charge for kernel software. Misc. past posts
mentioning scheduling and resource management
https://www.garlic.com/~lynn/subtopic.html#fairshare
for a few years there was "base" (free) operating system (kernel) offerings with optional "charged-for" kernel software (that was growing in size) ... until the cut-over was made to charge for all kernel software (and kernel product packaging collapsed back to single offering). About this time there was transition to "object code only" ... even tho software had started being charge-for with 23jun69 announcement, source was still available (for some products shipped with full source with source change/maintenance being done standard feature). "object code only" eliminated source availability (as standard option).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Mainframe upgrade done with wire cutters? Newsgroups: alt.folklore.computers Date: Sun, 23 Jan 2011 11:52:17 -0500re:
a side-effect of the gov. & unbundling ... was there were rules that prices charged had to be a profit. basically there was original development costs (upfront) plus production and support costs. some amount of the pre-unbundling development costs were grandfathered ... so development costs were primarily an issue with new products after unbundling.
some parts of the company weren't use to such accounting and had rather lavish processing. another characteristic was that some number of products were assumed to be price sensitive ... so there was "forecast" done for "low", "middle", and "high" price. basically
(price) * (related forecast) > (development costs) + (production & support costs)assumption was that "low" price would have larger forecast than "high" price ... and "low" price might even result in larger total revenue/profit.
however, some "new" products (done with much more lavish circumstances) found that there was no price that would result in enough customers to have a profit. however, there was some latitude in interpreting the rules.
in the mid-70s, there was a new "networking" product for the favorite son operating system for which there was no price-forecast that resulted in profit. however, there had been a vm370 "networking" product that the company was refusing to release (kill vm370 strategy) ... which had been developed with compareably "no resources" ... and had large forecast (if it were allowed to be announced) at numerous price points. The way forward for the favorite son operating system network was to have a "combined" product announcement ... allowing combining the development costs and combining the forecasts (for both products) ... which resulted in a way for the "mainstream" product to be announced (the combined product financials show a profit).
in the early 80s, the interpretation of the rules seemed to get even more relaxed. It was then sufficient to have totally different products in the same development organization ... the total revenue from all the products covered the total costs of developing & supporting all the products (in one case a 2-3 person product which had the same revenue as a couple hundred person product, allowing revenue from a "non-strategic" product to underwrite a product considered "strategic").
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Mainframe upgrade done with wire cutters? Newsgroups: alt.folklore.computers Date: Sun, 23 Jan 2011 13:41:37 -0500"Phxbrd" <lesliesethhammond@yahoo.com> writes:
after having spent a number of years redoing pieces of os/360 and
rewriting a lot of cp/67 as undergraduate ... I went to standard
interview job fair and was given the corporate programmer aptitude test
... which I apparently didn't pass ... but the science center hired me
anyway. misc. past posts mentioning the science center
https://www.garlic.com/~lynn/subtopic.html#545tech
the person doing the job interviews (at the univ) was from the san jose
plant site ... and he couldn't understand why I was being given a job
offer ... wasn't even entry position ... started me out straight out of
school at higher level ... of course, besides doing all the stuff at the
univ ... I had also been called in to help setup some of the early BCS
operation ... recent reference:
https://www.garlic.com/~lynn/2010q.html#59 Boeing Plant 2 ... End of an Era
somebody made a comment about the above post ... that when they went to work for lockheed ... they were told that the "real" winner of the C5A competition was the 747 (aka having lost, boeing turned it into a much more successful/profitable commercial airplane).
also as mentioned in the Boeing ref above ... for quite some time, I thot that the renton datacenter was the largest around ... but recently one of the Boyd (I had sponsored Boyd's briefings at IBM) biographies mentioned in 1970, he had done a stint running "spook base" which was a $2.5B "windfall" for IBM (approx. ten times the value of mainframes in the renton datacenter) ... inflation adjustment is about factor of seven ... or approx. $17.5B in today's dollars.
Marketing people on Boeing account ... claims it was a salesman on the Boeing account that motivated the corporation's change from straight commission to "quotas". Supposedly the day 360 was 1st announced, Boeing walked in to the salesman's office with large 360 order (and knew significantly more about what 360 was, than the salesman). The commission was much larger than the CEO's compensation ... resulting in the company creating the sales "quota" plan for the following year. The following year, Boeing walked in with another large order ... resulting in the salesman exceeding 100% of quota before the end of Jan. The salesman then left and formed a large computer services company.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Mainframe upgrade done with wire cutters? Newsgroups: alt.folklore.computers Date: Sun, 23 Jan 2011 17:33:00 -0500greymausg writes:
there was period when 386dx (w/fpu) had line cut & packaged with 16bit external bus (instead of 32bit) ... and sold cheaper as 386sx (w/o fpu) ... sort of like 8088 version of 8086.
that summer overseas builders had built up a big inventory of 286
machines for xmas seasonal sales. the 386sx just blew them (the 286
machines) out of the water and the 286 machines were sold at deep
discount that fall (effectively able to drop 386sx into 286
motherboard/machine and get higher performance)
https://en.wikipedia.org/wiki/Intel_80386
the 370 equivalent was that the incremental cost of producing a 370/158 dropped so low (comparable to some old figures regarding incremental manufacturing costs for producing high volume autos) ... that it became possible to sell 3031 for the same or less than 370/158.
the 370/158 had microcode engine that was shared between the 370 cpu function and the integrated channel function. for 303x, they created a "channel director" ... which was a 158 microcode engine w/o the cpu microcode and just the integrated channel function ... coupled with a 3031 processor ... which was 158 microcode engine w/o the integrated channel microcode and just the cpu microcode (in theory, a "single processor" 3031 actually had two 158 engines rather than just one).
old posts with some 1990 prices discussion about 286, 386, 486 (after
the 88 xmas sesson where market droped out of the 286 market)
https://www.garlic.com/~lynn/2001n.html#79 a.f.c history checkup... (was What specifications will the standard year 2001 PC have?)
https://www.garlic.com/~lynn/2001n.html#80 a.f.c history checkup... (was What specifications will the standard year 2001 PC have?)
https://www.garlic.com/~lynn/2001n.html#81 a.f.c history checkup... (was What specifications will the standard year 2001 PC have?)
https://www.garlic.com/~lynn/2001n.html#82 a.f.c history checkup... (was What specifications will the standard year 2001 PC have?)
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Rare Apple I computer sells for $216,000 in London Newsgroups: alt.folklore.computers, aus.electronics, aus.computers Date: Sun, 23 Jan 2011 18:14:42 -0500Seebs <usenet-nospam@seebs.net> writes:
almost exactly a year before morris worm (nov88)
https://en.wikipedia.org/wiki/Morris_worm
the xmas exec is basically social engineering ... distributing a compromised executable and getting people to load & execute.
this is slightly different from convention for automatic execution. that grew up with various office applications that evolved on local, private, safe, closed business networks. this infrastructure was then transferred to the wild anarchy of the internet w/o the necessary safety and countermeasures (aka just reading an email could result in automatic execution)
bitnet (along with EARN in europe) was higher education network (significantly
underwritten by IBM and using similar technology that was used for the
corporate internal network) ... past posts mentioning bitnet &/or earn
https://www.garlic.com/~lynn/subnetwork.html#bitnet
some old email by person charged with setting up EARN:
https://www.garlic.com/~lynn/2001h.html#email840320
the internal network was larger than the arpanet/internet from just
about the beginning until possibly late '85 or early '86. misc.
past posts mentioning internal network
https://www.garlic.com/~lynn/subnetwork.html#internalnet
I was blamed for online computer conferencing on the internal network in the late 70s and early 80s. The folklore is that when the executive committee was told about online computer conferencing (and the internal network), 5of6 wanted to fire me.
Later, somewhat as a result, a research was paid to study how I
communicated ... got copies of all my incoming & outgoing email, logs of
all my instant messages, sat in the back of my office for nine months
taking notes face-to-face and phone conversations (sometimes went with
me to meetings). This also turned into stanford phd thesis and material
for some number of papers and books. misc. past posts
https://www.garlic.com/~lynn/subnetwork.html#cmc
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Rare Apple I computer sells for $216,000 in London Newsgroups: alt.folklore.computers, aus.electronics, aus.computers Date: Sun, 23 Jan 2011 23:06:59 -0500Anne & Lynn Wheeler <lynn@garlic.com> writes:
bitnet annoucement on vmshare
http://vm.marist.edu/~vmshare/browse.cgi?fn=BITNET&ft=MEMO
tymshare made its vm370/cms online computer conferencing available to
SHARE user group organization in aug76
http://vm.marist.edu/~vmshare/
recent post about the internal network ...
https://www.garlic.com/~lynn/2011.html#4
including old email about plans to convert the internal
network to sna/vtam
https://www.garlic.com/~lynn/2011.html#email870306
also references old email about the executive committee being told that
PROFS was an SNA application (among other things) used to justify
converting the internal network to sna/vtam:
https://www.garlic.com/~lynn/2006x.html#email870302
in this old post
https://www.garlic.com/~lynn/2006x.html#7
and somewhat similar discussion here ... where somebody forwarded
me a lengthy log of email discussing how sna/vtam could be
the nsfnet backbone
https://www.garlic.com/~lynn/2006w.html#email870109
in this old post
https://www.garlic.com/~lynn/2006w.html#21
some of the same people involved in the above referenced email exchanges
(about sna/vtam for nsfnet backbone) ... were later involved in the
transfer of cluster scale-up ... mentioned in this old post about jan92
meeting in ellison's conference room:
https://www.garlic.com/~lynn/95.html#13
also referenced in this other email
https://www.garlic.com/~lynn/lhwemail.html#medusa
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 25 Jan, 2011 Subject: Credit cards with a proximity wifi chip can be as safe as walking around with your credit card number on a poster Blog: LinkedInlots of RFID work has been done for EPC ... basically next generation barcode for inventory applications. Various efforts have used the technology to encode static data ... like what is on payment card magstripe ... and use them in contactless environments. Other is iso14443 which has gotten a lot of play in transit turnstyle applications. In the late 90s, some from the transit industry asked if a x9.59 financial transaction could be done within the power & elapsed time turnstlye requirements, using chip less expensive than transit but much more secure than "secure" fiancial contact chips.
card security code was countermeasure to white card account number
guessing attack (generate magstripe info from formula) ... hash the
rest of the information encode with bin/bank secret.
https://en.wikipedia.org/wiki/Card_security_code
however, at least the early 90s, skimming attacks copied valid mastripe for creating counterfeit. x9.59 standard was from the x9a10 financial standard working group which had been given the requirement to preserve the integrity of the financial infrastructure for ALL retail payments
nacha did pilot close to x9.59. old ref from wayback machine:
https://web.archive.org/web/20070706004855/http://internetcouncil.nacha.org/News/news.html
earlier response to the RFI
https://www.garlic.com/~lynn/nacharfi.htm
there were a number of "safe" payment products being pushed at the start of the century that got high marks from large merchants ... until they were told that there would effectively be a surcharge on top of highest fee they were already paying (severe cognitive dissonance after decades of indoctrination that much of interchange fee was proportional to fraud).
US institutions have had 40% (for some large instiutitons 60%) of their bottom line coming from payments, as compared to less than 10% for Euroepean institutions. As a result US institutions have been a lot more resistant to any kind of disruptive change (merchants had expected a significant drop in interchange fees for the "safe" products).
in addition to card security wiki ... there is magstripe wiki (which
has had major rewrite recently, since the old version references Los
Gatos lab where I had several offices and labs):
https://en.wikipedia.org/wiki/Magnetic_stripe_card
The magstripe wiki references the IBMer that invented magstripe ... but the old verson also mentioned that magstripe standard was run out of los gatos lab for first decade or so ... also gatos lab did some of the early ATM (cash) machines
In addition to x9.59 standard, I did chip design ... that could be
used contactless and (in combination with x9.59) could be evesdropped
on live transactions and crooks couldn't still use information for
fraudulent transactions (also countermeasure to crooks using info
from data breaches for fraudulent transactions). I gave presentation
at Intel Developer Forum in the trusted computing track ... mentioning
the chip was at least as secure as TPM chip and 1/10th (or less) the
cost.
https://web.archive.org/web/20011109072807/http://www.intel94.com/idf/spr2001/sessiondescription.asp?id=stp%2bs13
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@GARLIC.COM (Anne & Lynn Wheeler) Subject: Re: Testing hardware RESERVE Newsgroups: bit.listserv.ibm-main Date: 25 Jan 2011 08:19:25 -0800PaulGBoulder@AIM.COM (Paul Gilmartin) writes:
above references System/R which was the original relational/SQL
done in bldg. 28
https://www.garlic.com/~lynn/submain.html#systemr
Another approach ... is a CKD channel program with compare&swap
semantics that was developed for HONE in the late 70s (US operation was
possibly largest single-system-image, loosely-coupled operation in the
world at the time) ... was more efficient than RESERVE/RELEASE (but not
as efficient as ACP RPQ) ... since it involved additional rotation. At
one-time there was extensive discussions with the JES2 multi-spool group
doing something similar. Misc. past posts mentioning internal HONE
system
https://www.garlic.com/~lynn/subtopic.html#hone
Later I needed an inverse of RESERVE for large cluster operation ... in recovery operation needing to remove a specific processor from the configuration ... I wanted a FCS switch operation to allow everybody ... but the processor that has been removed from the configuration (there is a failure mode where a processor stops, appears to have failed, and then later resumes ... potentially just before doing some I/O operation with global impact, aka it doesn't realize that it has been removed from the configuration).
One of the problems was that FCS was being quite distracted with the complex effort to layer FICON on top of it (somewhat in the manner that ALL current day CKD is done by simulation on top of underlying FBA).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Rare Apple I computer sells for $216,000 in London Newsgroups: alt.folklore.computers, aus.electronics, aus.computers Date: Tue, 25 Jan 2011 13:52:35 -0500Morten Reistad <first@last.name> writes:
regarding Melinda's pages with some mainframe historic documents moving
... there was some comment that princeton was removing her pages because
of possible hate crimes issues ... over her comments about MVS.
https://www.leeandmelindavarian.com/Melinda#VMHist
she had a multi-file postscript version that was many tens of megabytes (with lots of pictures) that I converted to PDF (4mbytes) and did an awz/kindle conversion. frequently kindle conversion becomes smaller file ... but with all the images, the kindle version is twice as large (9mbytes, as pdf).
other of the PDF files with figures that are line-drawings using characters didn't convert nearly as well (and converted to smaller files) for kindle ... with the characters in the drawings being "flowed".
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@GARLIC.COM (Anne & Lynn Wheeler) Subject: Re: Long-running jobs, PDS, and DISP=SHR Newsgroups: bit.listserv.ibm-main Date: 25 Jan 2011 11:02:30 -0800shmuel+ibm-main@PATRIOT.NET (Shmuel Metz , Seymour J.) writes:
The issue was if power was lost in the middle of write operation ... the channel/controller could continue to transfer data ... basically filling in with zeros (data was no longer coming from processor memory). The disk write could complete ... with the propagated zeros and correct error correcting (ECC) information written (based on the propagated zeros).
On restoration of power ... there would be no disk error reading the record ... just that record would be all zeros from the point power was lost.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: History of copy on write Newsgroups: comp.arch Date: Tue, 25 Jan 2011 16:53:19 -0500timcaffrey@aol.com (Tim McCaffrey) writes:
multics was done w/pli ... and had none of the buffer overflow problems
of unix. old post
https://www.garlic.com/~lynn/2002l.html#42 Thirty Years Later: Lessons from the Multics Security Evaluation
with references to
http://www.acsac.org/2002/papers/classic-multics.pdf
and
http://csrc.nist.gov/publications/history/karg74.pdf
and a reference to virtual machine work done on the 4th flr:
https://web.archive.org/web/20090117083033/http://www.nsa.gov/research/selinux/list-archive/0409/8362.shtml
part of the unix issue was that string/array conventions in C makes it almost as hard to *NOT* shoot yourself in the foot ... as it is in many other environments to actually shoot yourself in the foot (conventions in many other environments result in having to work really hard to have buffer overflows ... even for some assembler environments where there are specific kinds of coding conventions).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 25 Jan, 2011 Subject: WikiLeaks' Wall Street Bombshell Blog: Financial Crime Risk, Fraud and Securityre:
BofA update: Bank of America braces itself for fallout from WikiLeaks
disclosures report says
http://www.computerworld.com/s/article/9203180/Bank_of_America_braces_itself_for_fallout_from_WikiLeaks_disclosures_report_says
and more general wikileak item
Ralph Nader: Wikileaks and the First Amendment
http://www.counterpunch.org/nader12212010.html
all the articles seem to draw the analogy with the pentagon papers ... somebody obtains the information (potentially anonymously) and provides it for publication. The Nadar article raises the question why all the obfuscation and misdirection that has been going on regarding wikileaks role in any publication.
wikileak related ... not particularly financial ... but gov. ... more along the lines of the old pentagon papers:
A Clear Danger to Free Speech
http://www.nytimes.com/2011/01/04/opinion/04stone.html?_r=1
latest, article also mentions BofA:
Assange vows to drop 'insurance' files on Rupert Murdoch
http://www.theregister.co.uk/2011/01/12/wikileaks_insurance_files/
related to attacks on wikileaks founder
US Wikileaks investigators can't link Assange to Manning
http://www.theregister.co.uk/2011/01/25/assange_cant_be_tied_to_manning_says_report
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Rare Apple I computer sells for $216,000 in London Newsgroups: alt.folklore.computers, aus.electronics, aus.computers Date: Wed, 26 Jan 2011 13:27:15 -0500Roland Hutchinson <my.spamtrap@verizon.net> writes:
1) it went directly from transport/level four to LAN/MAC ... bypassing network/level three ... violating OSI model
2) it supporting "internetworking" ... a non-existent layer in the OSI model (approx. between transport/networking)
3) it went directly to LAN/MAC interface ... a non-existent interface in the OSI model (sitting approx. in the middle of layer 3 networking).
one of the other differences between ISO and IETF that has been periodically highlighted is that IETF (aka internet standards) requires that interoperable (different) implementations be demonstrated before progressing in the standards process. ISO can pass standards for things that have never been implemented (and potentially are impossible to implement).
misc. past posts mentioning HSP, ISO, OSI, etc
https://www.garlic.com/~lynn/subnetwork.html#xtphsp
note that fed. gov. in the late 80s was mandating that internet be eliminated and replaced with ISO (aka GOSIP).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: Melinda Varian's history page move Blog: IBM Historic ComputingMelinda Varian's history page move
Some of info about 70s & 80s was somewhat garbled since it was only based on information available externally.
For instance in the wake of FS failure and the mad rush to get products back into the 370 product pipeline, the favorite son operating system did managed to convince the corporation to kill vm370, shutdown the burlington development group, and move all the people to POK to support MVS/XA development (or otherwise it wouldn't meet its "ship" schedule). There was a "VMTOOL" done as part of MVS/XA development ... but it was never intended to be shipped to customers (eventually coming out as a "migration aid".
Endicott did manage to save the vm370 product mission, but it had to reconstitute a development group from scratch.
The corporation was planning on delaying the announcement of the burlington closing until the last possible minute ... to minimize the potential that members might find alternatives to moving to POK. When the information was leaked early ... there was a serious hunt to find the source. During those last weeks the halls in burlington was very furtive people ... nobody wanted to be seen talking to anybody else (or otherwise being suspected as the source of the leak). There was a minor joke that head of POK was a significant contributor to (DEC) VAX/VMS ... because of some number of burlington people that managed to leak away to DEC.
Other minor note was some claim that the 3rd or 4th internal CMSBACK
release being the first (before morphing into workstation datasave,
ADSM and current TSM). Reference to original internal CMSBACK release
was done in the late 70s ... which then went through a number of
internal releases
https://www.garlic.com/~lynn/lhwemail.html#cmsback
note: VMTOOL (internal only virtual machine for supporting mvs/xa development) is different from VMTOOLS (internal VM related online conferencing using TOOLSRUN).
VMTOOL was suppose to be an internal-only mvs/xa development tool ... it was never suppose to be released to customers. originally virtual machine product was to be killed off ... until endicott saved the product mission and endicott had to recreate a product group from scratch. The effects of endicott having to recreate a product group from scratch can sort of be implied from Melinda's history comments about the extremely poor product quality (during endicotts startup phase).
Melinda's history also mentions that VM/XA Migration Aid was based on
the VM Tool built for use by the MVS/XA developers Even though the vm
tool was never intended to be released to customers ... there was
eventually a case made that MVS customers needed it as an aid for
migrating from MVS to MVS/XA. ... small extract:
Although the button IBM was handing out said that the Migration Aid
was more than just a migration aid, it was not very much more. Release
1 was intended to be used to run MVS/SP in production while running
MVS/XA in test. All virtual machines were uniprocessors, for
example. VM/SP guests were also supported, and RSCS and PVM would run,
but CMS was supported only for maintaining the system. However, the
group in Poughkeepsie was working hard to produce a real VM/XA
system.
... snip ...
also see the preceding paragraph to above (this info was just what was available externally).
trivia ... who leaked to burlington that the corporation had decided to kill off vm370, shutdown the burlington group, and move all the people to POK as part of supporting MVS/XA (POK had made the case that they wouldn't be able to meet the MVS/XA ship schedule w/o those resources).
I would contend that one of the reasons that CMS/XA development was being done in YKT was exactly because the there had been no plan for a vm/cms product (and weren't staffed or organized for such a product).
.. some core of the burlington resources moved to POK were preserved for the vmtool work ... but none of the other skills/resources were kept for what they had been doing in burlington (product support, cms, etc). That was why the vmtool release as the (mvs to mvs/xa) migration aid .... had nothing else done for xa-mode ... just what had been done for the internal MVS/XA development.
Also see the SIE discussion which points to old email discussing the
difference between (original) 3081 SIE and 3090 SIE. The 3081 SIE had
never been intended for high-performance production work with high
frequency invokation. In 3081, there was limited microcode space
... so part of invoking the SIE instruction was "paged" microcode
... done by the service processor (uc.5) that was paging from 3310/FBA
https://www.garlic.com/~lynn/2011.html#62 SIE - CompArch
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: A brief history of CMS/XA, part 1 Blog: IBM Historic Computingfrom long ago and far away
and from a week earlier
Date: 06/18/84 11:51:10
From: wheeler
re: pam changes; fyi ... updates described in ciajoint to split the
dchsect into two blocks are being picked up by the cms/xa group,
pretty much as is ... and will support the aligning the data block
on 4kedf page boundaries. Still negotiating over the issues of picking
up full pam support.
... snip ... top of post, old email index
In the above, somebody had taken a bunch of internal stuff and included them as part of a proposed joint study with CIA.
PAM ... full paged mapped filesystem support ... I had originally done
in cp67 and migrated to vm370. DCSS (primarily changes to CMS code to
be included in shared segments) for vm370 release 3 ... was very small
subset of PAM ... but reworked to work with DMKSNT. Recent references
to DMKSNT:
https://www.garlic.com/~lynn/2011.html#21 zLinux OR Linux on zEnterprise Blade Extension???
https://www.garlic.com/~lynn/2011.html#28 Personal histories and IBM computing
https://www.garlic.com/~lynn/2011.html#74 shared code, was Speed of Old Hard Disks - adcons
lots of past mentions of page mapped filesystem
https://www.garlic.com/~lynn/submain.html#mmap
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: A brief history of CMS/XA, part 1 Blog: IBM Historic Computingre:
One of the issues regarding running out of 16mbyte virtual address space (mentioned in cms/xa requirements) is slightly analogous to MVS running out of application 16mbyte virtual address space (i.e. because of extensive pointer API, MVS kernel image occupied 8mbytes of every address space ... and "common segment" was threatening to take 6mbytes at large installations (concurrently running applications needed their own unique locations in the common segment for each subsystem).
The original PAM/shared-segment implementation allowed shared segments to appear at different locations. In the vm370 release 3 DCSS very small subset using DMKSNT ... each application "shared segment" was given a unique location in a hypothetical installation "global" address space (to avoid users having application address conflicts). Some installations then went to multiple hypothetical "global" address spaces ... multiple different copies of a shared system application fixed at different virtual addresses (hoping that a user could find some combination of the shared systems that they wanted, which wouldn't have virtual address space conflicts).
In the "dynamic" flavor, available contiguous virtual address would be
chosen dynamically at load time for each cms user (possibility that
same shared system appearing simultaneously at different virtual
address for different cms users). Misc. past posts mentioning issues
with dynamic virtual location
https://www.garlic.com/~lynn/submain.html#adcon
these recent posts mentioning the MVS common segment/CSA issue
https://www.garlic.com/~lynn/2011.html#45 CKD DASD
https://www.garlic.com/~lynn/2011.html#79 Speed of Old Hard Disks - adcons
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: New-home sales in 2010 fall to lowest in 47 years Blog: FacebookNew-home sales in 2010 fall to lowest in 47 years
from above:
Sales for all of 2010 totaled 321,000, a drop of 14.4 percent from the
375,000 homes sold in 2009, the Commerce Department said Wednesday. It
was the fifth consecutive year that sales have declined after hitting
record highs for the five previous years when the housing market was
booming.
... snip ...
and the news spin on business tv: DEC2010 saw "surge" in new-home sales (fine print compared to nov2010, nov2010 was the lowest ever, dec2010 was 2nd lowest, but still a "surge")
it occurred to me that it might be a version of pump&dump ... helping
with the hype to get dow to close above 12k. Cramer did interview a
couple yrs ago claiming that it was widespread practice to take a
position in the market and then flurry of slanted news (& while
illegal, nobody was worried because SEC was too dumb to figure it
out). recent refs to above:
https://www.garlic.com/~lynn/2010h.html#41 Profiling of fraudsters
https://www.garlic.com/~lynn/2010p.html#43 WikiLeaks' Wall Street Bombshell
there is slightly different view ... in the madoff hearings there was
testimony by a person that had tried for a decade to get SEC to do
something about Madoff. In the weeks around the hearings, the person
wouldn't appear in public and/or give interviews. When pressed, a
spokesperson made some reference to believing that the only reason
that SEC didn't do anything for a decade about Madoff, was because it
was heavily under influence of criminal organizations (capable of
extreme violence, who might object to his efforts to bring down
Madoff). a few recent refs to above:
https://www.garlic.com/~lynn/2010q.html#21 Ernst & Young called to account -- should Audit firms be investigated for their role in the crisis?
https://www.garlic.com/~lynn/2010q.html#40 Ernst & Young sued for fraud over Lehman
https://www.garlic.com/~lynn/2011.html#46 What do you think about fraud prevention in the governments?
https://www.garlic.com/~lynn/2011.html#48 What do you think about fraud prevention in the governments?
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: What do you think about fraud prevention in the governments? Blog: Financial Crime Risk, Fraud and SecurityCrisis Panel Report Pins Blame on Wall Street, Washington
the dissenting positions seem to ignore many of the items from this thread
including the part that rating agencies played in allowing unregulated loan originators to unload everything they wrote at triple-A ratings (so they no longer had to care about loan quality and/or borrower's qualifications)
... some archive from the thread:
https://www.garlic.com/~lynn/2011.html#46 What do you think about fraud prevention in the governments?
https://www.garlic.com/~lynn/2011.html#48 What do you think about fraud prevention in the governments?
https://www.garlic.com/~lynn/2011.html#49 What do you think about fraud prevention in the governments?
https://www.garlic.com/~lynn/2011.html#50 What do you think about fraud prevention in the governments?
https://www.garlic.com/~lynn/2011.html#53 What do you think about fraud prevention in the governments?
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: A brief history of CMS/XA, part 1 Blog: IBM Historic Computingre:
... for a little more fun, later in the day on 26jun84:
Date: 06/26/84 19:31:17
To: HONE branch office id (se on large customer account)
Cc: <*guess who*>
From: wheeler
hear anything more from pok/kingston??? I noticed <*guess who*>
forwarded the vmshare cms/xa article (that i sent out) to
pok/kingston. maybe that will hurry things along.
... snip ... top of post, old email index, HONE email
I had an ongoing process with Tymshare where they sent me regular tape dump of all the VMSHARE (and later added PCSHARE) online computer conferencing files ... which I then made available internally.
vmshare archive:
http://vm.marist.edu/~vmshare/
above reference's Melinda's (princeton) home page (which has yet to disappear).
misc. past posts mentioning HONE
https://www.garlic.com/~lynn/subtopic.html#hone
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: IBM S/360 Green Card high quality scan Blog: IBM Historic Computingre:
an (another 26jun84) item regarding IPCS replacement in REXX
Date: 06/26/84 15:11:37
To: endicott organization
From: wheeler
re: dumprx;
dumprx is over three years old. ... unfortunately I've made several
presentations on it at share & other groups over two years ago. I'm
just getting messages that non-IBM company is coming out with
something that is supposedly very similar to dumprx & implementation.
DUMPRX started as a project to demonstrate that system programming
could be done in higher level language ... and specifically that some
number of IBM system programming projects currently done in assembler
would be appropriately done in a higher level procedural language that
was interpreted ... i.e. demonstrate REXX feasibility.
Well, three years has come and gone ... it looks like somebody else is
releasing it ... even tho IBM ain't interested.
One of the primary issue of DUMPRX was given the concept about how to
implemented it ... several other things fell out. As a result my total
effort todate has been under six weeks ... which should be comperable
for anybody attempting to duplicate the effort. Anybody starting from
DUMPRX concept and putting some real effort (more than six weeks)
could/have turn out much more functional product.
... snip ... top of post, old email index
at the time DUMPRX was extensively used by internal datacenters and
majority of VM PSRs (internal product support people). misc.
past posts mentioning dumprx
https://www.garlic.com/~lynn/submain.html#dumprx
when rexx was still very young (not a product and still called "rex"), I wanted to demonstrate that it wasn't just another pretty scripting language. The objective was to implement a replacement for IPCS (written in assembler) in REXX working half-time over period of three months (and it would have ten times the function and be ten times faster).
The initial objective was achieved in a couple weeks ... so the rest of the time was spent doing "automated" debugging scripts that would perform most of the typical debugging operations a real person might do ... as well as scan of most storage looking for anomalies and specific failure signatures (being ten times faster than IPCS assembler was important here). DUMPRX session could be saved & restarted ... so human could resume an automated session where extensive analysis had already been performed.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 27 Jan 2011 Subject: Melinda Varian's history page move Blog: IBM Historic Computingre:
Note that CMS/XA is being started well after the start for MVS/XA (and
shutdown of burlington development group) ... and not being available
until 370/xa has been at customers for several years (another
indication that originally there were no plans for VM/XA product)
Date: 10/26/82 13:45:07
To: wheeler
Lynn,
Just logged on this ID and found your messages and notes about
development 'processes'. I have to agree with all of the comments I
read. It is MOST frustrating to have to work in this environment - as
I'm sure you appreciate.
Last week, *XXXXX* gave his 'state of the nation' address to the
assembled VM group here in POK. I was dismayed at some of the things
he said; for example - some VM work is being sub-contracted (sorry,
'vendored') to *iiiiiiiiiii*. Of course, the work they do is "simple",
not the advanced type which we do "here", but it costs him 1 million
dollars per year for this 'simple' work (and it is done by 8
programmers). Similarly, CMS/XA is being sub-contracted to Yorktown,
but again, this is "simple work". He didn't mention the fact that a 6
or 8 man group is contemplating generating something like 60,000 lines
of code by mid-1985, which works out at at least 500 LOC per person
per month - 5 or 10 times what is optimistically expected here in POK
CP development.
If only we weren't all bogged down in this mire of 'process' which
management has decided is vital, and if, as you propose, a decent
amount of time was allocated to education (the right kind, of course),
I'm sure we wouldn't be in the position we are now.
It almost seems we are getting back to the stage immediately
preceeding the Tandem memos again, but what can be done??????? I wish
I knew!
... snip ... top of post, old email index
I was blamed for online computer conferencing on the internal network
in the late 70s and early 80s ... part of which came to be referred to
as Tandem Memos. There was article on the phenomena in nov81
datamation. The folklore is that when the executive committee was
informed of online computer conferencing (and the internal network),
5of6 wanted to fire me.
Date: 3 October 1984, 14:45:53 EDT
From: <*guess who*>
To: large distribution
The Turbo CMS (CMS/XA) IPFS is now available for your reading
pleasure. If you want a copy (if you are really going to read it --
it is 160 pages long) send me your mailing address & I will send you a
copy. Those of you here in Yorktown can stop by 89-N15 & get a copy.
Please send all comments to me. Thanks,
<*guess who*>
... snip ... top of post, old email index
other recent cms/xa reference:
https://www.garlic.com/~lynn/2011b.html#19 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#20 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#23 A brief history of CMS/XA, part 1
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Rare Apple I computer sells for $216,000 in London Newsgroups: alt.folklore.computers Date: Thu, 27 Jan 2011 22:13:33 -0500Peter Flass <Peter_Flass@Yahoo.com> writes:
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 28 Jan, 2011 Subject: The Zippo Lighter theory of the financial crisis (or, who do we want to blame?) Blog: Financial CryptographyThe Zippo Lighter theory of the financial crisis (or, who do we want to blame?)
slightly shorter version of my earlier comments ... in
What caused the financial crisis. (Laying bare the end of banking.)
http://financialcryptography.com/mt/archives/001268.html
turning loans/mortgages into toxic CDOs created huge new set of transactions for wallstreet being able to take possibly 15-20% in fees and commissions
reference to possibly $27T in triple-A rated toxic CDOs transactions
done during the bubble (possibly $5.4T for wallstreet?)
Evil Wall Street Exports Boomed With 'Fools' Born to Buy Debt
https://www.bloomberg.com/news/articles/2008-10-27/evil-wall-street-exports-boomed-with-fools-born-to-buy-debt
there are reports that the industry tripled in size (as percent of GDP) during the bubble. also NY comptroller reported that wallstreet bonuses spiked more than 400% during the bubble.
mortgages had been packaged as CDOs during the S&L crisis (obfuscate the underlying value); but w/o the triple-A ratings had very little market.
this time unregulated loan orginators found that they could pay for triple-A ratings (even when both the originators and rating agencies knew they weren't worth triple-A ... from the congressional rating agency hearings, fall2008) ... which exploded the market for the toxic CDOs ... and eliminating any reason for the loan orginators to care about loan quality or borrowers qualifications. Since they could immediately unload everything at triple-A ... they just turned into loan/mortgage mills ... their revenue only limited by how many & how fast they could turn over the loans.
no-documentation, no-down, 1% interest-only payment ARMs found an exploding market among speculators; with real-estate inflation running 20-30% ... it met possibly 2000% ROI (flipping before the rates adjusted ... further churning transactions and boosting inflation). These mortgages become the equivalent of the '20s "Brokers' Loans" ... allowing the real-estate market to be turned into the equivalent of the '20s stock market.
Individual compensation on the loan originator and the wall street sides (of packaging loans as triple-A rated toxic CDOs) was so great that it overrode any possible concern as to the effects on the institution, the economy, and/or the country (with being able to buy triple-A ratings sitting in the middle).
The triple-A rating made the toxic CDOs appear acceptable to large
number of institutions that wouldn't otherwise deal in the
instruments. At the end of 2008, the four too-big-to-fail regulated,
"safe" institutions were carrying $5.2T of the toxic CDOs
"off-balance" (courtesy of repeal of Glass-Steagall and their
"unregulated", risky investment banking arms). ref
Bank's Hidden Junk Menaces $1 Trillion Purge
>http://www.bloomberg.com/apps/news?pid=newsarchive&sid=akv_p6LBNIdw&refer=home
GSE lost a major portion of the mortgage market share as mortgages were being packaged up as triple-A rated toxic CDOs and sold in these other markets.
Earlier in 2008, a number of toxic CDO sales ... totaling several tens of billions had gone for 22cents on the dollar. If the too-big-to-fail institutions had to bring their off-balance toxic CDOs back onto the balance sheet, they would have been declared insolvent and forced to be liquidated. TARP funds had supposedly been to buy up these toxic assets but they obviously didn't know the magnitude of the problem (the amount appropriated wouldn't have made a dent in the problem).
It took more than a year of legal efforts to get Federal Reserve to disclose what it has been doing; buried in the information is reference to them buying up these assets at 98cents on the dollar.
now this is archaic, long-winded post from 1999 discussing several of
the problems ... including the fact that in 1989 CITI was nearly taken
down by its ARM mortgage portfolio (it got out of the business and
required a private bailout to stay in business)
https://www.garlic.com/~lynn/aepay3.htm#riskm
role forward and CITI is holding the largest share of the $5.2T in triple-A toxic CDOs (effectively mostly an ARM mortgage portfolio) and requires another bailout to stay in business. Repeal of Glass-Steagall didn't directly cause the problem, repeal of Glass-Steagall just allowed several of these too-big-to-fail institutions to enormously help fuel the (triple-A rated toxic CDO) loan/mortgage mill and side-step the regulations designed to keep them out of risky behavior.
related to the ZIPPO theme ... regulations had kept the hotspots of greed and corruption separate and damped down ... then the period of extremely lax regulation and de-regulation ... allowed the greed and corruption hotspots to merge and create a financial firestorm.
one of the suggestions has been to RICO wallstreet
https://en.wikipedia.org/wiki/Racketeer_Influenced_and_Corrupt_Organizations_Act
for three times the $27T involved in the triple-A rated toxic
CDOs ($81T).
Evil Wall Street Exports Boomed With 'Fools' Born to Buy Debt
https://www.bloomberg.com/news/articles/2008-10-27/evil-wall-street-exports-boomed-with-fools-born-to-buy-debt
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Mainframe upgrade done with wire cutters? Newsgroups: alt.folklore.computers Date: Fri, 28 Jan 2011 16:04:11 -0500re:
... faulty memory and not carefully rereading the referenced wiki article ... clearly says that 387 wasn't ready at the time of 386(dx) and so early motherboards had slots for 287.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 28 Jan 2011 Subject: A brief history of CMS/XA, part 1 Blog: IBM Historic Computingre:
the theoretical cutoff for this discussion group is 1985 ... a lot of the vm/xa and cms/xa stuff is 85 or later.
CMSSTRUCT MEMO from vmshare touches on some issues (from 28Feb82):
http://vm.marist.edu/~vmshare/browse.cgi?fn=CMSTRUCT&ft=MEMO
the referenced TSS CSECT & PSECT comes into play regarding my posts
about shared segments being able to dynamically appear at multiple
different locations simultaneously
https://www.garlic.com/~lynn/submain.html#adcon
The pascal and semi-privilege address space is somewhat related to
rewrite I did of cp spool system in pascal and ran it in a virtual
address space. recent post:
https://www.garlic.com/~lynn/2010k.html#26 Was VM ever used as an exokernel?
https://www.garlic.com/~lynn/2010k.html#35 Was VM ever used as an exokernel?
in the following, note the emphasis on *intelligent workstations*. one
of the issues regarding arpanet/internet passing the internal network
in size was ability for workstations to be network nodes ... while the
corporate communication group was restricting PCs and workstations to
terminal emulation. some past posts
https://www.garlic.com/~lynn/subnetwork.html#emulation
Date: 20 April 1983, 17:18:11 EST
From: <*guess who*>
About a dozen SEs met this past Monday and Tuesday to discuss CMS and
CMS/XA requirements. Here is my summary of some of the important
issues that came out of that meeting.
Summary of input from the CMS Specialists Focus Session
One big incompatible change is much better than many small changes.
Substantial improvements in CMS are desired more than compatibility
with what we have today. A competitive product that has
incompatibilities can be sold more easily that a completely compatible
product that is not competitive.
File sharing is the number one requirement. It is more important than
SNA support. It is more important than XA support. Any "new" or
"improved" CMS is not new or improved if it lacks file sharing. A new
file system cannot be sold if it does not have file sharing. (At
least, file sharing must be announced as "on the way" when the new
file system is announced. Skilled marketing could hold the customers
off for a few months.) Customers would accept a system that is up to *
slower if it has file sharing. The level of integrity provided by
EDF today is considered adequate.
Having only one CMS is very important. This single CMS (at least the
user and application programming interfaces) must run across the
product line from intelligent workstations to 308x machines. Function
may be subsetted on the workstation CMS, but the interface should be
consistent.
A well defined interface is very important. It is essential if one is
to provide single CMS user and programming interfaces across a range
of machines. The interface described in the CMS Restructure proposal
(opaque, machine like, well defined) was received with enthusiasm, and
was described as being well worth the conversion effort.
Intelligent workstations are going to be as important as XA in the
latter half of this decade, if not more so. Great growth in
workstation usage is expected. No one seems really sure how they
relate to and interface with the mainframes.
Multitasking and window systems are seen as important functions if CMS
is to be a competitive system. They are especially important when
workstations are considered. Flexibility is very important in a
window system, as there will be many different workstations out there.
OS ATTACH is important for applications running in service virtual
machines. It is much less important for applications running in the
CMS user's machine.
DOS simulation is little used. Most users compile DOS programs by
submitting them to service machines running VSE as guest operating
systems. No complaints if DOS simulation is done away with. Except
that VSAM is widely used. OS VSAM is preferred, but it does not
support FBA devices.
<*guess who*>
... snip ... top of post, old email index
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Fri, 28 Jan 2011 22:08:36 -0500Charles Richmond <frizzle@tx.rr.com> writes:
The person that i believe did that port ... left a couple yrs later ... after having done an educational program for his home atari. when nobody at ibm was interested, he offered it directly to atari. atari then contacted ibm ... who unleashed the lawyers. he departed ibm within 24 hrs.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Sat, 29 Jan 2011 10:24:45 -0500Peter Dassow <z80eu@arcor.de> writes:
this was internal network 1978 ... i was getting a copy for vm370/cms from TYMSHARE. TYMSHARE appeared to have gotten copy from Stanford SAIL pdp10 ... for their pdp10 and then ported to vm370/cms and made executable available on their commercial online vm370/cms timesharing service (there is folklore that when TYMSHARE executives 1st heard that games were on their system, they insisted all games be removed ... until they were informed that games were accounting for 1/3rd of their online revenue).
TYMSHARE had also made their online computer conferencing available
to SHARE as VMSHARE in aug76 ... VMSHARE archive here:
http://vm.marist.edu/~vmshare/
I was setting things up to get regular copy of the VMSHARE files to make them available internally ... so was periodically dropping by TYMSHARE offices for that reason ... and other reasons.
I finally got a copy and made executable version available on the
internal network ... the internal network was larger than the
arpanet/internet from just about the beginning until possibly
late '85 or early '86. misc. past posts mentioning internal network
https://www.garlic.com/~lynn/subnetwork.html#internalnet
TYMSHARE wiki reference:
https://en.wikipedia.org/wiki/Tymshare
random TYMSHARE trivia ... GNOSIS was a 370-based operating system
developed by TYMSAHRE ... as part of M/D purchase in '84 ... i was
brought in to do audit of GNOSIS
https://en.wikipedia.org/wiki/GNOSIS
as part of its spinoff to keylogic (for keykos)
https://en.wikipedia.org/wiki/KeyKOS
... i still have gnosis hardcopy someplace in boxes.
Doug was also working for TYMSHARE at the time ... and there was some
concern what would happen in the M/D purchase ... so I set up some
interviews trying to get him hired
https://en.wikipedia.org/wiki/Douglas_Engelbart
for other drift ... misc. past posts mentioning adams website:
https://www.garlic.com/~lynn/2005u.html#25 Fast action games on System/360+?
https://www.garlic.com/~lynn/2006y.html#18 The History of Computer Role-Playing Games
https://www.garlic.com/~lynn/2006y.html#19 The History of Computer Role-Playing Games
https://www.garlic.com/~lynn/2007g.html#0 10 worst PCs
https://www.garlic.com/~lynn/2010q.html#70 VMSHARE Archives
note ... the tcp/ip is the technology basis for the modern internet,
NSFNET backbone was the operational basis for the modern internet, and
CIX was the business basis for the modern internet. in the 80s I was
working with NSF and some of the expected locations on what was to
become the NSFNET backbone ... an old email reference:
Date: 11/14/85 09:33:21
From: wheeler
re: cp internals class;
I'm not sure about 3 days solid ... and/or how useful it might be all
at once ... but I might be able to do a couple of half days here and
there when I'm in washington for other reasons. I'm there (Alexandria)
next tues, weds, & some of thursday.
I expect ... when the NSF joint study for the super computer center
network gets signed ... i'll be down there more.
BTW, I'm looking for a IBM 370 processor in the wash. DC area
running VM where I might be able to get a couple of userids and
install some hardware to connect to a satellite earth station & drive
PVM & RSCS networking. It would connect into the internal IBM pilot
... and possibly also the NSF supercomputer pilot.
... snip ... top of post, old email index, NSFNET email
however, various internal politics prevented me being able to bid on
the NSFNET backbone RFP. The director of NSF attempted to help by
writing the company a letter 3Apr1986, NSF Director to IBM Chief
Scientist and IBM Senior VP and director of Research, copying IBM CEO)
... but that just aggravated the internal politics (there were
comments about what we already had running was at least five years
ahead of all NSFNET bid submissions to build something
new). . misc. other old NSFNET email
https://www.garlic.com/~lynn/subnetwork.html#nsfnet
recent post about various internal operations were claiming
SNA/VTAM for NSFNET:
https://www.garlic.com/~lynn/2011b.html#10 Rare Apple I computer sells for $216,000 in London
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Sat, 29 Jan 2011 10:37:44 -0500re:
for other archaic nsfnet topic drift
Date: 09/30/85 17:27:27
From: wheeler
re: channel attach box; fyi;
I'm meeting with NSF on weds. to negotiate joint project which will
install HSDT as backbone network to tie together all super-computer
centers ... and probably some number of others as well. Discussions
are pretty well along ... they have signed confidentiality agreements
and such.
For one piece of it, I would like to be able to use the cambridge
channel attach box.
I'll be up in Milford a week from weds. to present the details of the
NSF project to ACIS management.
... snip ... top of post, old email index, NSFNET email
I had an internal HSDT (high-speed data transport) project ... had
terrestrial links as well as TDMA earth stations with transponder on
SBS4. I was involved in doing various kinds of things ... including
numerous projects using NSC's HYPERchannel ... misc. past posts
mentioning HSDT
https://www.garlic.com/~lynn/subnetwork.html#hsdt
and little more HSDT drift
https://www.garlic.com/~lynn/2011.html#92 HELP: I need a Printer Terminal
Date: 11/16/85 14:54:06
From: wheeler
re: fireberd; fyi, i've got the beginnings of a turbo pascal programs
that I hope will eventually be able to ship fireberd data to the host.
Right now I have two programs, one that supports two ASYNCH ports and
acts as an ascii terminal display (ultramux) with one asynch port and
as the fireberd printer using COM2. I also have another Turbo program
that interfaces to xxxxx's high level language interface for PC327x
and will logon to VM and perform some other simple tasks. I hope to
merge the two so that I can log both fireberd error data and ultramux
output on a vm host.
... snip ... top of post, old email index
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 29 Jan 2011 Subject: A brief history of CMS/XA, part 1 Blog: IBM Historic Computingre:
and for quite a bit CMS/XA topic drift:
misc. past posts mentioning HSDT (high-speed data transport) project
and/or (NSC) HYPERchannel
https://www.garlic.com/~lynn/subnetwork.html#hsdt
Date: 28 November 1983, 11:19:35 EST
To: wheeler
I am from IBM-xxxxxx, presently on assignment in Bob O'Hara's
department working on CMS/XA. I spent my last 2 years in Endicott in
the CMS development group. I was reading your paper "Historical Views
of VM Performance" and in the bibliography found a reference to
"VM/370 Hyperchannel Support I" which you have not published
yet. Prior to my coming to the US, I was in the "field" working for
the French Atomic Energy Commission (roughly equivalent to the SLAC)
where I was in charge of VM and VTAM. They had a large heterogeneous
network involving IBM and CDC computers and at that time (that was in
79-80) they got several NSC hyperchannels for which they developed
support in JES2. I worked in putting JES2 under CMS and that's the way
we've got VM into the network.
Would it be possible to get a copy of your paper ; is there any plan
to include your work in the product ? I am interested in any
information related to VM - NSC. Thank you.
... snip ... top of post, old email index
some of this came up in thread yesterday about ADVENTURE on vm370/cms
in late 70s:
https://www.garlic.com/~lynn/2011b.html#30 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#31 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#32 Colossal Cave Adventure in PL/I
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Sun, 30 Jan 2011 09:50:26 -0500Anne & Lynn Wheeler <lynn@garlic.com> writes:
I was never actually allowed to sign such a contract and/or install HSDT
... later the network to tie together all the super-computer centers
became the NSFNET backbone (which then evolved into the modern internet)
... after the original NSFNET backbone RFP. misc. past email related to
"NSFNET"
https://www.garlic.com/~lynn/lhwemail.html#nsfnet
misc. past posts mentioning HSDT
https://www.garlic.com/~lynn/subnetwork.html#hsdt
The NSFNET Backbone RFP went out specifying "T1" links (in part because HSDT was already running T1 backbone). However, the winning response (for $11.2M) actually only installed 440kbit links. Then sort of to meet the letter of the RFP ... they installed T1 trunks with telco multiplexors running multiple 440kbit links. Sarcastically made comments that possibly they could claim a "T5" network ... since there was possibility that some of the T1 trunks were in turn multiplexed over T5 trunks at some point. For various & sundry reasons there is old folklore that actual resources that went into NSFNET backbone was three to four times the RFP $11.2M.
old email reference to director of NSF sending corporate letter (copying
CEO) regarding HSDT (hoping to help with the internal politics, but
it apparently just made it worse):
https://www.garlic.com/~lynn/2006s.html#email860417
old email regarding internal politics attempting to position sna/vtam
for nsfnet backbone
https://www.garlic.com/~lynn/2006w.html#email870109
old email reference to NSFNET RFP finally being awarded
https://www.garlic.com/~lynn/2000e.html#email880104
... later possibly believing that they could blunt my sarcastic remarks, for the NSFNET backbone T1->T3 upgrade RFP response, I was asked to be the redteam (the blueteam was something like 2-3 dozen people from half dozen or so labs around the world). At the final review, I presented 1st, then a few minutes into the blueteam response, the executive running the review pounded on the table and said he would lie down in front of a garbage truck before he let any but the blueteam response go forward (apparently even executives could understand my response was vastly superior). I got up and left ... there were even a few others that walked out with me.
past posts in this thread:
https://www.garlic.com/~lynn/2011b.html#30 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#31 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#32 Colossal Cave Adventure in PL/I
above also posted to linkedin ietf discussion:
http://www.linkedin.com/groups/Today-IETF-Turns-25-Slashdot-83669.S.40470495?qid=d0b1ba86-6e72-48b0-9694-d382555fcdda
http://www.linkedin.com/blink?msgID=I78973238_20&redirect=leo%3A%2F%2Fplh%2Fhttp%253A*3*3lnkd%252Ein*3tqm3Pj%2F4ium&trk=plh
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Sun, 30 Jan 2011 16:36:14 -0500Peter Flass <Peter_Flass@Yahoo.com> writes:
references:
VMCM0350 -- PL/I port of WOOD0350 for VM/CMS
http://risc.ua.edu/pub/games/colossal/colossal.zip
http://risc.ua.edu/pub/games/cms/colossal.vmarc
but the above server no longer exists
however, there is the wayback machine:
https://web.archive.org/web/*/http://risc.ua.edu/pub/games/cms/colossal.vmarc
but at the moment it is saying try again later.
note adventure wiki (and several other references that appear to repeat what is in the wiki) ... says adventure became available on ibm mainframes on vm/cms in late 1978 ustilizing pl/i version. tymshare had just taken fortran version from pdp10 to vm/cms (and apparently had gotten the pdp10 version from stanford just a couple miles away).
while waiting for copy from Tymshare ... spring '78 email requesting copy
https://www.garlic.com/~lynn/2008s.html#email780321
response a couple weeks later from UK (corporate site near/at
univ. location)
https://www.garlic.com/~lynn/2006y.html#email780405
and my response shortly later same day
https://www.garlic.com/~lynn/2006y.html#email780405b
adventure related email a few days later
https://www.garlic.com/~lynn/2007m.html#email780414
and another email a month later mentioning pli version
https://www.garlic.com/~lynn/2007m.html#email780517
past posts in this thread:
https://www.garlic.com/~lynn/2011b.html#30 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#31 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#32 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#34 Colossal Cave Adventure in PL/I
past posts in longer thread from year ago.
https://www.garlic.com/~lynn/2010d.html#57 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#64 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#65 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#67 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#68 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#74 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#75 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#77 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#82 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010d.html#84 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010e.html#4 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010e.html#9 Adventure - Or Colossal Cave Adventure
and a similar thread from four yrs ago:
https://www.garlic.com/~lynn/2007o.html#8 Original Colossal Cave Adventure
https://www.garlic.com/~lynn/2007o.html#11 Original Colossal Cave Adventure
https://www.garlic.com/~lynn/2007o.html#15 "Atuan" - Colossal Cave in APL?
note while melinda's princeton site is moving ... previous ref:
https://www.garlic.com/~lynn/2011b.html#4 Rare Apple I computer sells for $216,000 in London
it still will live on at the wayback machine (aka www.princeton.edu) ...
note a previous incarnation of the Melinda's webpages at
"pucc.princeton.edu" used to have a copy of zork source for vm/cms ...
while it has been gone for some time ... it also still lives on at the
wayback machine:
https://web.archive.org/web/20010124044900/http://pucc.princeton.edu/~melinda/
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 30 Jan, 2011 Subject: Internal Fraud and Dollar Losses Blog: Financial Crime Risk, Fraud and SecurityInternal Fraud and Dollar Losses
from above:
Research Suggests Banks Don't Catch Most Internal Fraud Schemes. A new
Aite research report proves internal fraud is more damaging than many
financial-services companies realize.
... snip ...
When we were doing what is now frequently called "electronic commerce", I tried to include a requirement that anybody that had any contact in anyway with webserver needed a FBI background check. Obviously that didn't fly.
The issue is that information from previous transaction (skimming, data breaches, evesdropping, etc) can be used by crooks to perform fraudulent financial transactions. Have tried using a number of metaphors regarding the problem; dual-use vulnerability (same information needed by crooks is also required in dozens of business processes at millions of locations around the world), security proportional to risk (value of information to merchant is profit from transaction possibly a few dollars, value of information to crook is account balance or credit limit ... hundreds or thousands of dollars; crooks can afford to outspend attacking by possibly a factor of 100 times as merchant can afford to spend defending).
In the mid-90s, possibly because of work on electronic commerce, I was
invited to participate in the x9a10 financial standard working group,
which had been given the requirement to preserve the integrity of
the financial infrastructure for all retail payments (i.e. ALL,
internet, point-of-sale, unattended, high value, low value, debit,
credit, gift card, ACH, wireless, contact, contactless, etc; aka
ALL). One of the features of the resulting financial transaction
standard was slight tweak to the current paradigm to eliminate
information from current transactions as a risk (didn't do anything
about skimming, evesdropping, data breaches, etc; it just
eliminating crooks being able to use the information for fraudulent
transactions).
https://www.garlic.com/~lynn/x959.html#x959
now the biggest use in the world today of SSL is this early thing we had done for electronic commerce to hide financial transaction details. However, the x9a10 work eliminated the need to hide such details ... so it also eliminates the major use of SSL in the world today.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: 1971PerformanceStudies - Typical OS/MFT 40/50/65s analysed Newsgroups: alt.folklore.computers Date: Mon, 31 Jan 2011 10:42:07 -0500old post ... with parts of presentation I gave at fall68 SHARE meeting
starting with sysgen for MFT11 (on 360/67 mostly running as 360/65), I would take the stage1 output (stage2) and create independent jobs for most job steps and re-arrange the order of lots of things. The objective was to 1) run stage2 in production jobstream and 2) order files & PDS members for optimal arm seek motion. For lot of typical univ. workload, it achieved nearly three times thruput improvement.
a slight improvement from IBM came with release 15/16 where it was possible to specify VTOC location. VTOC was the highest used disk location and always on cylinder 0. Then careful placement of files and PDS would have highest used next to VTOC and decreasing used data at increasing cylinder locations. Starting with release 15/16, it was possible to place VTOC in the middle of the disk and arrange high use data on both sides of VTOC.
I had also done some work rewriting significant portions of cp67 to reduce pathlength ... the presentation also lists some numbers of that pathlength reduction.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Mon, 31 Jan 2011 15:12:46 -0500re:
as per referenced thread here in afc from 2010:
https://www.garlic.com/~lynn/2010d.html#64 Adventure - Or Colossal Cave Adventure
I was sent Advent_CMS.ZIP (2008) which contains complete set of source and executable files (along with necessary data).
and as also referenced, in ibm-main discussion there was reference to
also one on file 269 on cbttape
http://www.cbttape.org/
other posts in this thread:
https://www.garlic.com/~lynn/2011b.html#30 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#31 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#32 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#34 Colossal Cave Adventure in PL/I
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: 1971PerformanceStudies - Typical OS/MFT 40/50/65s analysed Newsgroups: alt.folklore.computers Date: Mon, 31 Jan 2011 19:35:08 -0500Peter Flass <Peter_Flass@Yahoo.com> writes:
kindle version up here:
https://www.leeandmelindavarian.com/Melinda/neuvm.azw
as mentioned recently in linkedin ibm historic computing group ... after
doing the kindle version ... and getting melinda to put up on her new
webpage ... i was rereading ... and found i had comments on many of the
pages ... misc. ibm historic computing post
https://www.garlic.com/~lynn/2011b.html#25 Melinda Varian's history page move
old email looking for the original cms source multi-level update
process.
https://www.garlic.com/~lynn/2006w.html#email850906
https://www.garlic.com/~lynn/2006w.html#email850906b
https://www.garlic.com/~lynn/2006w.html#email850908
it was implemented in EXEC that iterated multiple times for project that added 370 virtual machine simulation to cp/67 (370 had new instructions and virtual memory hardware tables were different from 360/67) ... aka "cp/67-h". Then the "cp/67-i" updates were changes to cp/67 to run on 370 (originally in cp/67 virtual machine).
i had lots of stuff archived from the early 70s and even some stuff from univ. in the 60s. these were replicated on multiple different tapes in the (same) almaden tape library. the above exchange was fortunately (shortly) before almaden had an operational problem where random tapes were mounted for "scratch" (managing to wipe out all my redundant copies).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Tue, 01 Feb 2011 09:59:36 -0500re:
for other "super-computer" topic drift ... old post with reference to
jan92 meeting in ellison's conference room
https://www.garlic.com/~lynn/95.html#13
old email mentioning "medusa"
https://www.garlic.com/~lynn/lhwemail.html#medusa
the last email reference in the above
https://www.garlic.com/~lynn/2006x.html#email920129
was possibly only hrs before effort was transferred and we were told we
couldn't work on anything with more than four processors. misc. past
posts mentioning ha/cmp
https://www.garlic.com/~lynn/subtopic.html#hacmp
Then came press on 2/17/92 with reference to being limited to
scientific and technical only (aka eliminating commercial)
https://www.garlic.com/~lynn/2001n.html#6000clusters1
and then 11May92, comment that they were caught by
surprise
https://www.garlic.com/~lynn/2001n.html#6000clusters2
being told we couldn't work on anything with more than four processors (as well as the earlier incident involving NSFNET backbone) contributed to our leaving later in '92.
Now two of the other people mentioned in the jan92 meeting in ellison's conference room, not long later show up at a small client/server startup responsible for something called the "commerce server". We were brought in to consult because they wanted to do payment transactions on the server (the small client/server startup had also invented this technology called "SSL" that they wanted to use). The result is now frequently called "electronic commerce".
part of the "electronic commerce" effort involved something called the
"payment gateway" which handled payment transaction flow between
"commerce servers" and the payment networks. misc. past posts
mentioning payment gateway
https://www.garlic.com/~lynn/subnetwork.html#gateway
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Tue, 01 Feb 2011 15:20:37 -0500Peter Flass <Peter_Flass@Yahoo.com> writes:
the version i got 2008 Advent.CMS.ZIP uses TREAD/TWRITE for
terminal input/output.
DCL TREAD ENTRY (CHAR(133),FIXED BIN(31),CHAR(133),
FIXED BIN(31),FIXED BIN(31)) OPTIONS (ASM INTER);
DCL TWRITE ENTRY (CHAR(133),FIXED BIN(31),FIXED BIN(31))
OPTIONS (ASM INTER);
the ZIP file includes a "wellput.asm" .. that includes the following
WELLPUT TITLE 'W E L L P U T -- WYLBUR/TSO I/O INTERFACE FROM PLI' 00010000 *********************************************************************** 00180000 • WELLPUT - * 00190000 • THE PURPOSE OF THIS MODULE IS TO SIMULATE THE I/O ROUTINES TREAD * 00200000 • AND TWRITE USED BY THE ADVENTURE GAME. * 00210000 • * 00220000 • CALLING SEQUENCES: * 00230000 • * 00240000 • TREAD (PROMPT_MESSAGE,PROMPT_LENGTH, MESSAGE_AREA,LENGTH,RTN_CODE) * 00250000 • * 00260000 • TWRITE (MESSAGE,MESSAGE LENGTH,RETURN CODE) * 00270000 • * 00280000 *********************************************************************** 00290000basically maps TREAD/TWRITE to TGET/TPUT (tso) assembler macros (which are simulated under CMS):
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 01 Feb, 2011 Subject: Productivity And Bubbles Blog: IBMersThe fall 2008 congressional hearings into the rating agencies had testimony that the rating agencies were selling triple-A ratings on toxic CDOs (when both the sellers and the rating agencies knew they weren't worth triple-A).
CDOs had been used in the S&L crisis to obfuscate the underlying mortgage values ... but w/o triple-A ratings they found little market. Being able to pay for triple-A ratings gave unregulated loan originators nearly unlimited supply of funds being able to immediately sell-off every loan they wrote w/o regard to loan quality or borrowers qualifications (the only limiting factor became how fast they could write the loans).
During the recent bubble/mess, estimate that $27T in triple-A rated
toxic CDO transactions were done ... w/o having to resort to federal
money.
Evil Wall Street Exports Boomed With 'Fools' Born to Buy Debt
https://www.bloomberg.com/news/articles/2008-10-27/evil-wall-street-exports-boomed-with-fools-born-to-buy-debt
Loan originators found big new market in real estate speculators. Real estate speculators found no-documentation, no-down, 1% interest only payment ARMs could return 2000% in real estate markets with 20-30% inflation (with the speculation further fueling inflation and transaction turn-over ... possibly new mortgage every yr on same properties). These loans (and the real estate market) became the equivalent of the "Brokers Loans" that were responsible for the the 20s stock market bubble & crash (from 30s Pecora Hearings into the '29 crash and also resulted in Glass-Steagall).
Obfuscating the real-estate bubble and crash was the repeal of
Glass-Steagall allowed unregulated risky investment banking arms of
safe&secure regulated depository institutions to buy lots of the
triple-A rated toxic CDOs and carry them off-balance. Estimate that at
the end of 2008, the four largest too-big-to-fail financial
institutions were carrying $5.2T "off balance"
Bank's Hidden Junk Menaces $1 Trillion Purge
>http://www.bloomberg.com/apps/news?pid=newsarchive&sid=akv_p6LBNIdw&refer=home
Earlier in 2008, several tens of billions of toxic CDOs had gone for 22cents on the dollar. If the $5.2T had been brought back on the books, the institutions would have been declared insolvent and had to be liquidated. On the front-end of the triple-A toxic CDO transactions were the huge speculation in the real-estate market, bubble and bust.
On the backend of the triple-A toxic CDO transactions was all the institutions holding the instruments and all the news trying to keep these too-big-to-fail institutions in business (shouldn't have been allowed to play in such risky business w/o the repeal of Glass-Steagall), was a big distraction from what was going on the real-estate bubble/bust side.
In the middle were all these new fees and commissions on this new way
of packaging loans as triple-A rated toxic CDOs. There was report that
the financial industry tripled in size (as percent of GDP) during the
bubble and the NY comptroller had report that the aggregate wallstreet
bonuses spiked over 400% during the bubble. Possibly aggregate of 15%
(new fees & commissions) on the $27T would be around $4T.
http://www.businessweek.com/stories/2008-03-19/the-feds-too-easy-on-wall-streetbusinessweek-business-news-stock-market-and-financial-advice
With the gov. leaning over backwards to keep the institutions in
business ... apparently a little thing like money laundering isn't a
big deal ... when DEA followed the money trail used to buy drug
smuggling planes:
http://www.bloomberg.com/news/2010-06-29/banks-financing-mexico-s-drug-cartels-admitted-in-wells-fargo-s-u-s-deal.html
and
http://www.huffingtonpost.com/zach-carter/megabanks-are-laundering_b_645885.html?show_comment_id=53702542
and
http://www.taipanpublishinggroup.com/tpg/taipan-daily/taipan-daily-080410.html
The personal compensation on wallstreet was so great that it easily
overrode any possible concern what the $27T in triple-A rated toxic
CDO transactions might do to the institutions, the economy and/or the
country. The business people were telling the risk dept. to fiddle the
inputs until they got the desired outputs (GIGO)
Evil Wall Street Exports Boomed With 'Fools' Born to Buy Debt
https://www.bloomberg.com/news/articles/2008-10-27/evil-wall-street-exports-boomed-with-fools-born-to-buy-debt
How Wall Street Lied to Its Computers
http://bits.blogs.nytimes.com/2008/09/18/how-wall-streets-quants-lied-to-their-computer
a lot of the talk about the computer models were wrong frequently has been obfuscation and misdirection.
Subprime = Triple-A ratings? or 'How to Lie with Statistics
https://web.archive.org/web/20071111031315/http://www.bloggingstocks.com/2007/07/25/subprime-triple-a-ratings-or-how-to-lie-with-statistics/
There was recent note that GSEs (fannie/freddie) were responsible for something like $5T of the $27T ... supposedly the GSEs came to the triple-A toxic CDOs (mortgage securitization) rather late and that is why their percent of the market had fallen (in part because the GSE standards were so high)
The GSEs aren't w/o their problems ... along the lines of some of the too-big-to-fail institutions, but on a much smaller scale.I think it was CBS had news program that mentioned at one point, freddie had more lobbiests on their roles than employees (they apparently tried to put everybody in washington, all former congressman and high level gov. bureaucrats on retainer; lots of lobbying money for a GSE ... but couldn't touch the amount of lobbying money wall street was spending). There was item from 2008, that Warren Buffett had been the largest Freddie shareholder in 2000/2001 ... but sold all his shares because of the GSEs accounting methods.. Freddie had been fined $400m for $10B in inflated statements and CEO was replaced ... but CEO was allowed to keep tens (hundred?) of millions in compensation.
However, lots of companies were doing something similar ... even after
SOX ... which put in all sorts of audit procedures and penalties for
public company fraudulent financial filings. However, possibly because
GAO didn't think SEC was doing anything (during the past decade), they
started doing reports of uptick in public company fraudulent financial
filings (even after SOX).
http://www.gao.gov/new.items/d061053r.pdf
and
https://www.gao.gov/products/gao-06-1079sp
so did SOX 1) have no effect on fraudulent financial filings, 2) encouraged fraudulent financial filings, or 3) if it wasn't for SOX, all financial filings would have been fraudulent (???).
The explanation was that executives boosted their bonuses from the fraudulent filings ... and even if the filings were later corrected, they still didn't have their bonuses corrected. Note also in the congressional Madoff hearings, one of the persons testifying, had tried unsuccessfully for a decade to try and get SEC to do something about Madoff (there was something about being worried that SEC hadn't done anything because it was heavily under the influence of criminal organizatons).
There was also something about that GSE CRA/subprime accounted for less than 10% of the $27T. During most of the bubble, the GSEs continued to do pretty much what they had been doing before the bubble. Real CRA/subprime hardly shows up as a small blip in the bubble. The bubble came from the whole rest of the market ... outside of CRA (and outside GSEs). You don't get 20-30% inflation in high-end real estate from anything remotely resembling CRA (which is where the majority of the $27T came from ... effectively by definition CRA properties have hard time accounting for even small percent of the $27T).
this has $10B in CRA done in 2001 with plans on having $500B total for
the decade
http://findarticles.com/p/articles/mi_m0EIN/is_2001_May_7/ai_74223918/
that comes to about 2% of the $27T. There has been sporadic references
to CRA as having contributed to the bubble ... but at 2% ... it is
possibly obfuscation and misdirection.
Evil Wall Street Exports Boomed With 'Fools' Born to Buy Debt
https://www.bloomberg.com/news/articles/2008-10-27/evil-wall-street-exports-boomed-with-fools-born-to-buy-debt
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 02 Feb, 2011 Subject: Productivity And Bubbles Blog: IBMersre:
Part of the difference was that the original statement was that GSE
CRA/subprime accounted for less than 10% of the mortgages (as opposed
to 10% of the $27T). The CRA/subprime mortgages were way down at the
low-end of the market ... so aggregate value of those mortgages was
much smaller percentage of the overall $27T (the enormous inflation
and bubble happened at other places in the market)
Evil Wall Street Exports Boomed With 'Fools' Born to Buy Debt
https://www.bloomberg.com/news/articles/2008-10-27/evil-wall-street-exports-boomed-with-fools-born-to-buy-debt
Somewhat like pulling out a single budget item that accounts for 2% of the $14T federal debt ($280B) and claim that is responsible for all of the federal woes (again obfuscation and misdirection).
One of the other issues (involving Buffett) in 2008 was the muni-bond market totally collapsed (investors realizing that if rating agencies were selling triple-A ratings for toxic CDOs ... how could any ratings be trusted). Buffett then stepped in and started offering "insurance" to get muni-bond market back up and running.
However, muni-bond market is also under pressure from collateral damage as part of the bubble collapse 1) real-estate bubble implosion cuts municipal revenue ... affecting their credit rating and 2) with all the speculation, there appeared to be more demand than there actually was, then builders built a lot of new development for the demand illusion, and municipalities sold bonds to put in utilities and streets for the new developments (when the new properties sold, the new revenue would cover the costs of the new facilities ... which has yet to happen).
Similarly, local banks have had collateral damage ... commercial builders borrowed to put in new strip malls for the new home developments. The new home developments aren't selling, the new strip malls don't have buyers/leases, the commercial builders have to default, and the local banks are collateral damage.
The speculators using no-documentation, no-down, 1% interest only payment ARMs treat the real-estate market like the 20s stock market (possibly 2000% ROI in regions with 20-30% real-estate inflation; these mortgages equivalent to the Brokers' Loans behind the 20s stock market boom/crash) ... the real-estate boom/crash has collateral damage that spreads out through much of the economy.
Early spring, 2009 I was asked to take the scan of 30s Senate Pecora hearings (scanned the previous fall at Boston public library and online at the wayback machine), HTML it with heavy cross-index and lots of references between what went on then and what happened this time. Apparently there was some expectation that the new congress had some appetite to do something ... however after doing quite a bit of work, I got a call saying it wouldn't be needed after all (no substantial difference between the new congress and congresses from the previous couple decades).
misc. past posts mentioning muni-bond market:
https://www.garlic.com/~lynn/2008j.html#9 dollar coins
https://www.garlic.com/~lynn/2008j.html#20 dollar coins
https://www.garlic.com/~lynn/2008j.html#23 dollar coins
https://www.garlic.com/~lynn/2008k.html#16 dollar coins
https://www.garlic.com/~lynn/2008k.html#23 dollar coins
https://www.garlic.com/~lynn/2008o.html#45 The human plague
https://www.garlic.com/~lynn/2008o.html#52 Why is sub-prime crisis of America called the sub-prime crisis?
https://www.garlic.com/~lynn/2008p.html#60 Did sub-prime cause the financial mess we are in?
https://www.garlic.com/~lynn/2008q.html#11 Blinkenlights
https://www.garlic.com/~lynn/2008q.html#20 How is Subprime crisis impacting other Industries?
https://www.garlic.com/~lynn/2009b.html#78 How to defeat new telemarketing tactic
https://www.garlic.com/~lynn/2009c.html#29 How to defeat new telemarketing tactic
https://www.garlic.com/~lynn/2009d.html#77 Who first mentioned Credit Crunch?
https://www.garlic.com/~lynn/2009e.html#8 The background reasons of Credit Crunch
https://www.garlic.com/~lynn/2009n.html#47 Opinions on the 'Unix Haters' Handbook'
https://www.garlic.com/~lynn/2010f.html#81 The 2010 Census
https://www.garlic.com/~lynn/2010l.html#53 Who is Really to Blame for the Financial Crisis?
https://www.garlic.com/~lynn/2010p.html#17 What banking is. (Essential for predicting the end of finance as we know it.)
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Colossal Cave Adventure in PL/I Newsgroups: alt.folklore.computers Date: Wed, 02 Feb 2011 11:55:41 -0500re:
the TREAD/TWRITE statements appears to Orvyl(/Wylbur?) system which predate TSO (tget/tput macros). Stanford was one of the univ. that got 360/67 for tss/360 ... and then when tss/360 ran into problems ... they did their own virtual memory systems (akin to MTS at univ of michigan) ... the Wylbur user interface/editor later being ported to OS/360 (and later MVS).
the implication was that the tread/twrite PLI version had originally been done for orvyl back in the 60s at stanford.
reference to (pdp/dec) sail (& adventure) at stanford
http://infolab.stanford.edu/pub/voy/museum/pictures/AIlab/SailFarewell.html
another reference to (pdp/dec) sail (& adventure) at stanford
http://www.stanford.edu/~learnest/sailaway.htm
wiki page:
https://en.wikipedia.org/wiki/Colossal_Cave_Adventure
the adventure translation from sail to orvyl at stanford would appear to be a naturual. The fortran version from stanford PDP machine to tymshare (nearby in silicon valley) PDP machine would have been straight-forward ... as well as tymshare porting fortran version from dec to vm/cms.
misc. past posts mentioning Orvyl
https://www.garlic.com/~lynn/2007g.html#31 Wylbur and Paging
https://www.garlic.com/~lynn/2007g.html#33 Wylbur and Paging
https://www.garlic.com/~lynn/2007m.html#62 nouns and adjectives
https://www.garlic.com/~lynn/2008h.html#78 Microsoft versus Digital Equipment Corporation
https://www.garlic.com/~lynn/2010e.html#79 history of RPG and other languages, was search engine history
https://www.garlic.com/~lynn/2010e.html#82 history of RPG and other languages, was search engine history
https://www.garlic.com/~lynn/2010j.html#67 Article says mainframe most cost-efficient platform
https://www.garlic.com/~lynn/2010k.html#11 TSO region size
https://www.garlic.com/~lynn/2010p.html#42 Which non-IBM software products (from ISVs) have been most significant to the mainframe's success?
https://www.garlic.com/~lynn/2011.html#6 IBM 360 display and Stanford Big Iron
https://www.garlic.com/~lynn/2011.html#73 Speed of Old Hard Disks - adcons
https://www.garlic.com/~lynn/2011.html#86 Utility of find single set bit instruction?
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 02 Feb, 2011 Subject: Productivity And Bubbles Blog: IBMersre:
long-winded early '99 post that discusses some number of the details
of the current situation ... including in '89, citi had figured out
that its ARM mortgage portfolio could take down the institution (even
regulated safety&soundness), unloaded the portfolio, got out of the
market, and required a (private/foreign) bailout to stay in business.
https://www.garlic.com/~lynn/aepay3.htm#riskm
Then, late '99, Bank Modernization ("GLBA") act passes. On the floor of congress the rhetoric was the main purpose of the bill was if you were already a (regulated, depository institution) bank, you got to remain a bank; if you weren't already a bank, you didn't get to become one (specifically calling out walmart and microsoft as examples). "GLBA" also repealed Glass-Steagall (out of the 30s Pecora hearings designed to prevent repeat of unregulated risky investment banks taking down safe&sound regulated depository institutions), and "opt-out" privacy sharing provisions (federal pre-emption of the cal. bill in progress that required "opt-in" for privacy sharing).
Roll foward into this century and a combination of all the regulatory repeal as well as lack of enformcement of regulations that did exist ... allowed a lot of isolated greed and corruption hot-spots to combined together into financial firestorm.
The unregulated loan originators were able to use paying for triple-A ratings and loan securitization (toxic CDOs) as nearly unlimited source of funds ... and also to write loans that didn't meet GSE/fanny/freddie standards ("standardless" loan market was new w/o a lot of compeitition ... since everybody else had been writing loans that met some minimum set of standards).
Wallstreet adored the $27T in new transactions ... since it was brand
new major source of fees and comissions. Courtesy of Glass-Steagall
... unregulated investment banking arms of safe/sound depository
institutions could play in the fees&commissions (otherwise prevented
by regulatory safety & soundness restrictions). The major
too-big-to-fail depository institutions didn't necessarily directly
cause the real-estate market to be treated as the '20s stock market
... but they provided a lot of the funds that fueled the bubble.
Evil Wall Street Exports Boomed With 'Fools' Born to Buy Debt
https://www.bloomberg.com/news/articles/2008-10-27/evil-wall-street-exports-boomed-with-fools-born-to-buy-debt
At end of 2008, four largest too-big-to-fail carrying $5.2T in
triple-A rated toxic CDOs "off-balance"
Bank's Hidden Junk Menaces $1 Trillion Purge
>http://www.bloomberg.com/apps/news?pid=newsarchive&sid=akv_p6LBNIdw&refer=home
Note the large triple-A rated toxic CDOs are effectively mostly a form of ARM mortgage portfolio and the largest holder is citi ... which required a new (gov) bailout to stay in business (institutional knowledge from '89 seemed to have evaporated).
Now there are periodic smokescreens attempting to ascribe the problems involving trillions of dollars were caused by problems that only involved billions of dollars (difference involving two to three orders of magnitude).
GLBA footnote ... TARP funds were a drop in the bucket fixing
wallstreet ... the big bailout for wallstreet too-big-to-fail was the
Federal Reserve for regulated depository institutions. They had been
fighting release of information ... but courts finally got some of the
information:
http://www.csmonitor.com/USA/2010/1201/Federal-Reserve-s-astounding-report-We-loaned-banks-trillions
including they are paying 98cents on the dollar for off-balance toxic assets (that had been going for 22cents on the dollar). However, all this was only to institutions that had bank charters. There were still a couple of large wallstreet investment banks that were in deep trouble ... and weren't "banks". Supposedly GLBA would have prevented them from becoming banks ... but federal reserve managed to give them bank charters anyway ... so they would also have access to federal reserve assistance.
And as to GLBA federal pre-emption of cal. financial institution "opt-in" privacy sharing. In the middle of last decade I was at an annual, national privacy conference at the Renaissance hotel in wash dc ... and they had a panel discussion with the FTC commissioners. Somebody in the back of the room said he was associated with financial industry "call centers" and he claimed that none of the people answering "opt-out" (privacy sharing) calls had any method to record information (GLBA allowed financial industry to use your privacy information unless they had record of your "opt-out" ... as opposed to cal. only allowing use of your privacy information if they had record of your "opt-in").
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 02 Feb, 2011 Subject: zLinux OR Linux on zEnterprise Blade Extension??? Blog: MainframeZonere:
repeat of old upthread reference to early Jan92 meeting in Ellison's
conference room regarding massive parallel
https://www.garlic.com/~lynn/95.html#13
misc. old email involving both commercial (including DBMS) as well as
technical and scientific cluster product
https://www.garlic.com/~lynn/subtopic.html#hacmp
and old email regarding work on cluster scale-up
https://www.garlic.com/~lynn/lhwemail.html#medusa
the last email in the above was possibly only hrs before the cluster
scale-up was transferred and we were told we couldn't work on anything
with more than four processors. A couple weeks later (2/17/92) there
is announcement that it is only for technical and scientific
https://www.garlic.com/~lynn/2001n.html#6000clusters1
and repeat of upthread reference to "Release No Software Before Its
Time" (parallel processing nearly two yrs ago):
https://www.garlic.com/~lynn/2009p.html#43
https://www.garlic.com/~lynn/2009p.html#46
we had worked with a number of RDBMS vendors that had "cluster" products for the old VAX/VMS cluster product ... and had a number of things to say about short-comings/bottlenecks in that implementation. So drawing on
(1) past (mainframe) loosely-coupled experience ... my wife did stint
in POK responsible for loosely-coupled (mainframe) architecture where
she developed Peer-Coupled Shared Data architecture ... misc. past
posts
https://www.garlic.com/~lynn/submain.html#shareddata
which saw little uptake until sysplex (help account for her not
staying long in the position)
(2) avoiding various VAX/VMS cluster deficiencies ... but emulating the API ... minimizing effort needed for RDBMS vendors to port from VAX/VMS cluster to HA/CMP cluster.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 02 Feb, 2011 Subject: A brief history of CMS/XA, part 1 Blog: IBM Historic Computingwith regard to FBA mentioned upthread
misc. past posts mentioning CKD & FBA
https://www.garlic.com/~lynn/submain.html#dasd
including references to being told that even if I gave POK favorite son operating system, completely integrated & tested FBA support, I still needed a $26M business case (to cover documentation & training). I wasn't able to use lifetime costs for business case, it had to be incremental new disk sales; possibly $200m-$300m sales to have profit to cover $26M (and then claim was FBA support would just result in customers buying same amount of FBA as CKD).
currently all CKD is emulation on real FBA.
one of the issues for MVS and explosion in (vm370) 4341 installs ... is that 3370 FBA was the "mid-range" disk technology (3380 was high-end CKD but no CKD mid-range). MVS could somewhat play when 4341 was used to replace earlier 370s with existing CKD. eventually they came out with 3375 for MVS mid-range (i.e. CKD emulation on top of 3370).
past posts in this thread:
https://www.garlic.com/~lynn/2011b.html#19 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#20 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#23 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#29 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#33 A brief history of CMS/XA, part 1
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Speed of Old Hard Disks Newsgroups: alt.folklore.computers Date: Wed, 02 Feb 2011 23:23:05 -0500re:
from long ago and far away ...
Date: 03/15/87 11:51:24 PST
From: wheeler
re: spring '85, processor cluster memos;
Technology is getting there. Couple weeks ago heard mention of
Fujitsu 3mip vm/pc 370 card available later this year. Recently
somebody mentioned that when ordering 700+mbyte, 3380-performance,
5.25 harddisk, the vendor asked him if he would prefer to wait 3
months for shipment of double density drives. Friday night somebody
mentioned introductory offer on a populated 4megbyte memory board for
$650.
Shouldn't be too long before there could be a board with 16meg memory
and 5mip 370 (somewhere in the $1k-$3k range).
1) with dirt cheap memory, systems have got to come up with new
innovative methods of using that memory effectively. For instance UNIX
pipes rather than DASD scratch files can avoid significant amounts of
DASD I/O.
2) systems leveraging price/performance of the large quantity mass
produced boards for computational complexes.
3) 3380 in 5.25 packaging sizes
4) 3380 frames housing hoards of 5.25 drives with control unit that
stripes data across multiple drives (4k FBA with 256bytes blocks on
multiple different drives). Data is recoverable even in cases of
multiple drive failures. More efficient and better recovery than DASD
mirroring. And/or lots of memory for full-track caching in the
control unit, super-large tracks (i.e. *8 or *16 real track multiple)
and potential of simultaneous data transfer on all drives (plus all
the recoverability, say 16*3mbyte=48mbyte aggregate data transfer).
What other ways are there to leverage some of the consumer electronic
business? Can a japanese compact disk player be turned into a cheap
fiber optic modem? and/or would it be cheaper to post-manufacter the
compact disk player into a modem than it would be to set up a separate
line to just do a modem (getting both a compact disk drive and a fiber
modem in single package ... using common reed-soloman error chip and
other components)?
... snip ... top of post, old email index
the '85 processor cluster memos were referenced in this
old post
https://www.garlic.com/~lynn/2004m.html#17 mainframe and microprocessor
misc. old email mentioning 801, iliad, romp, risc, etc
https://www.garlic.com/~lynn/lhwemail.html#801
old posts mentioning getting to play disk engineer in bldgs. 14&15
https://www.garlic.com/~lynn/subtopic.html#disk
this is old email with special head that could read/write 16+2
tracks simultaneously ... somewhat similar data rate:
https://www.garlic.com/~lynn/2006s.html#email871230
in the late 70s, one of the engineers in bldg.14 got patent on what
would later be called RAID. It was eventually used by s/38 (because
disk failure was so tramatic in s/38 environment). recent ref
https://www.garlic.com/~lynn/2011.html#14 IBM Future System
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 03 Feb, 2011 Subject: vm/370 3081 Blog: IBM Historic Computingfrom long ago and far away ...
past tales about VM/SP1 changes significantly increased (vm370) kernel
multiprocessor overhead ... changes primarily targeted at improving
virtual ACP/TPF thruput. 308x was initially announced with
multiprocessor only but ACP/TPF didn't have multiprocessor
support. VM/SP1 multiprocessor support was reworked attempting to
increase the concurrency of overlapping virtual machine simulation
functions (like I/O) with virtual machine execution (specifically
targeted reducing ACP/TPF elapsed time by increased overlap
w/multiprocessor). However, this restructuring increased the overall
multiprocessor overhead (for special case ACP/TPF concurrency
... attempting to make use of 2nd otherwise idle processor in a
dedicated ACP/TPF environment) for all customers
https://www.garlic.com/~lynn/2007.html#email820512 .
https://www.garlic.com/~lynn/2007.html#email820513 .
https://www.garlic.com/~lynn/2007.html#email860219
a reference to 3081 being left-over, very expensive FS technology
http://www.jfsowa.com/computer/memo125.htm
wiki page:
https://en.wikipedia.org/wiki/IBM_3081
eventually 3083 was created motivated by ACP/TPF market ... basically
3081 with one processor removed. A slight problem was simplest was
remove processor-1 ... leaving processor-0 ... but processor-0 was
located at the top of the 3081, which would have left 3083 potentially
dangerously top-heavy.
https://www.garlic.com/~lynn/2009r.html#70 While watching Biography about Bill Gates on CNBC last Night
https://www.garlic.com/~lynn/2010d.html#79 LPARs: More or Less?
https://www.garlic.com/~lynn/2010e.html#23 Item on TPF
https://www.garlic.com/~lynn/2010i.html#78 IBM to announce new MF's this year
https://www.garlic.com/~lynn/2010n.html#16 Sabre Talk Information?
https://www.garlic.com/~lynn/2011.html#19 zLinux OR Linux on zEnterprise Blade Extension???
other posts mentioning 3083
https://www.garlic.com/~lynn/99.html#103 IBM 9020 computers used by FAA (was Re: EPO stories (was: HELP IT'S HOT!!!!!))
https://www.garlic.com/~lynn/2000b.html#65 oddly portable machines
https://www.garlic.com/~lynn/2000d.html#9 4341 was "Is a VAX a mainframe?"
https://www.garlic.com/~lynn/2000f.html#69 TSS ancient history, was X86 ultimate CISC? designs)
https://www.garlic.com/~lynn/2001b.html#37 John Mashey's greatest hits
https://www.garlic.com/~lynn/2001c.html#13 LINUS for S/390
https://www.garlic.com/~lynn/2001j.html#17 I hate Compaq
https://www.garlic.com/~lynn/2002c.html#9 IBM Doesn't Make Small MP's Anymore
https://www.garlic.com/~lynn/2002i.html#83 HONE
https://www.garlic.com/~lynn/2002m.html#67 Tweaking old computers?
https://www.garlic.com/~lynn/2002o.html#28 TPF
https://www.garlic.com/~lynn/2002p.html#58 AMP vs SMP
https://www.garlic.com/~lynn/2003g.html#30 One Processor is bad?
https://www.garlic.com/~lynn/2003p.html#45 Saturation Design Point
https://www.garlic.com/~lynn/2004.html#7 Dyadic
https://www.garlic.com/~lynn/2004c.html#35 Computer-oriented license plates
https://www.garlic.com/~lynn/2004e.html#44 Infiniband - practicalities for small clusters
https://www.garlic.com/~lynn/2004h.html#8 CCD technology
https://www.garlic.com/~lynn/2005.html#22 The Soul of Barb's New Machine (was Re: creat)
https://www.garlic.com/~lynn/2005j.html#16 Performance and Capacity Planning
https://www.garlic.com/~lynn/2005m.html#55 54 Processors?
https://www.garlic.com/~lynn/2005o.html#44 Intel engineer discusses their dual-core design
https://www.garlic.com/~lynn/2005s.html#7 Performance of zOS guest
https://www.garlic.com/~lynn/2005s.html#38 MVCIN instruction
https://www.garlic.com/~lynn/2006d.html#5 IBM 610 workstation computer
https://www.garlic.com/~lynn/2006l.html#30 One or two CPUs - the pros & cons
https://www.garlic.com/~lynn/2006m.html#32 Old Hashing Routine
https://www.garlic.com/~lynn/2006n.html#16 On the 370/165 and the 360/85
https://www.garlic.com/~lynn/2007.html#33 Just another example of mainframe costs
https://www.garlic.com/~lynn/2007.html#44 vm/sp1
https://www.garlic.com/~lynn/2007g.html#16 What's a CPU second?
https://www.garlic.com/~lynn/2007o.html#37 Each CPU usage
https://www.garlic.com/~lynn/2007t.html#30 What do YOU call the # sign?
https://www.garlic.com/~lynn/2008c.html#83 CPU time differences for the same job
https://www.garlic.com/~lynn/2008e.html#40 Fantasy-Land_Hierarchal_NUMA_Memory-Model_on_Vertical
https://www.garlic.com/~lynn/2008g.html#14 Was CMS multi-tasking?
https://www.garlic.com/~lynn/2008i.html#38 American Airlines
https://www.garlic.com/~lynn/2008i.html#50 Microsoft versus Digital Equipment Corporation
https://www.garlic.com/~lynn/2008i.html#51 Microsoft versus Digital Equipment Corporation
https://www.garlic.com/~lynn/2008i.html#57 Microsoft versus Digital Equipment Corporation
https://www.garlic.com/~lynn/2008p.html#27 Father Of Financial Dataprocessing
https://www.garlic.com/~lynn/2009g.html#66 Mainframe articles
https://www.garlic.com/~lynn/2009g.html#68 IT Infrastructure Slideshow: The IBM Mainframe: 50 Years of Big Iron Innovation
https://www.garlic.com/~lynn/2009g.html#70 Mainframe articles
https://www.garlic.com/~lynn/2009h.html#77 Operating Systems for Virtual Machines
https://www.garlic.com/~lynn/2009l.html#55 IBM halves mainframe Linux engine prices
https://www.garlic.com/~lynn/2009l.html#65 ACP, One of the Oldest Open Source Apps
https://www.garlic.com/~lynn/2009l.html#67 ACP, One of the Oldest Open Source Apps
https://www.garlic.com/~lynn/2009m.html#39 ACP, One of the Oldest Open Source Apps
https://www.garlic.com/~lynn/2010.html#1 DEC-10 SOS Editor Intra-Line Editing
https://www.garlic.com/~lynn/2010.html#21 Happy DEC-10 Day
https://www.garlic.com/~lynn/2010d.html#14 Happy DEC-10 Day
https://www.garlic.com/~lynn/2010i.html#24 Program Work Method Question
https://www.garlic.com/~lynn/2010j.html#20 Personal use z/OS machines was Re: Multiprise 3k for personal Use?
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Speed of Old Hard Disks Newsgroups: alt.folklore.computers Date: Fri, 04 Feb 2011 10:18:54 -0500re:
old email mentioning processor clusters, they wanted me in ykt all week on the subject ... but I was also scheduled to be in DC with NSF.
I was being pressured to doing the processor cluster all week and
getting somebody else to handle the NSF high-speed networking
presentation
https://www.garlic.com/~lynn/2007d.html#email850315
also mentioned here
https://www.garlic.com/~lynn/2011b.html#34 Colossal Cave Adventure in PL/I
https://www.garlic.com/~lynn/2011b.html#40 Colossal Cave Adventure in PL/I
Date: 03/14/85 12:25:05
From: wheeler
going to be here on monday? going to dc for nsf meeting on tuesday?
I figure I can take red-eye to NY monday night & get a flight to DC in
time for the nsf meeting. can be back in ykt some time on weds
did i blow enuf smoke on processor clusters?
... snip ... top of post, old email index, NSFNET email
... also leading up to the above:
Date: 03/13/85 10:32:58
From: wheeler
you've haven't said much recently .. i'll be in ykt at least next
thursday ... if not weds. also. I'll be in wash. dc on tuesday for NSF
pitch in the afternoon.
... snip ... top of post, old email index, NSFNET email
I had to get somebody else to do the pitch to NSF director so I could
do the processor cluster meetings in YKT ... the day before that
Date: 03/12/85 12:51:18
From: wheeler
Univ. of Cal. has gotten a $120m NSF grant to set-up a super computer
regional center in San Diego. I've been asked to make a pitch to them
at Berkeley ... getting lots of data into & out of the center
... apparently this work is beyond the leading edge of anything else
out there.
As a political move, in the past two days, a presentation was set-up
with the head of NSF for next Tuesday so he can get the pitch before
UofC sees it.
... snip ... top of post, old email index, NSFNET email
older network information
Date: 09/16/83 10:44:22
From: wheeler
re: csnet/arpanet; Arpanet is getting very restrictive about
connection ... just about have to have DoD contract. CSNET started up
with a NSF grant and has a large number of universities and other
locations have connection. Quite a few of the CSNET locations have
Arpanet gateway connections which provide mail forwarding facilities.
Both SJR & YKT have connections to CSNET machines which will transport
mail across the boundary. In addition, there is an external VNET
network called BITNET which currently has (at least) 30-40 locations.
YKT & CAMB. have connections to BITNET.
Implemented IBM security procedures currently require only
"authorized" nodeid/userid use of the gateways (i.e. the IBM gateway
machines will only process mail being sent by/to authorized
and registered nodeid/userid). Other mail is returned unsent.
... snip ... top of post, old email index
old reference to original SJR connection to CSNET
https://www.garlic.com/~lynn/internet.htm
including
https://www.garlic.com/~lynn/internet.htm#email821021
other old email mentioning NSFNET
https://www.garlic.com/~lynn/lhwemail.html#nsfnet
later processor cluster email
https://www.garlic.com/~lynn/lhwemail.html#medusa
as part of ha/cmp cluster scale-up
https://www.garlic.com/~lynn/subtopic.html#hacmp
misc. high speed networking (HSDT) posts
https://www.garlic.com/~lynn/subnetwork.html#hsdt
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Speed of Old Hard Disks Newsgroups: alt.folklore.computers Date: Fri, 04 Feb 2011 11:12:22 -0500Anne & Lynn Wheeler <lynn@garlic.com> writes:
some folklore at the time was that everybody figured that the supercomputer would go to berkeley ... but supposedly the UofC regents had schedule that UCSD would get the next new bldg. ... and some big portion of the NSF supercomputer grant would go for constructing a new bldg ... and so the regents decided that the supercomputer center had to be done in San Diego.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 04 Feb 2011 Subject: A brief history of CMS/XA, part 1 Blog: IBM Historic Computingother bits & pieces from long ago and far away
and
Date: 05/17/84 12:40:44
From: wheeler
there is cms/xa meeting around palo alto next week on the 22nd & 23rd
including meetings with INTEL, LSI, and Signetics.
... snip ... top of post, old email index
and
Date: 05/17/84 12:42:31
From: wheeler
re: cms/xa; i assume that message was invitation ... I accept and am
planning on attending. It might be worthwhile to also invite
xxxx-xxxxx from the ACIS as an observer. I know it isn't directly the
subject of the meetings ... but I have a feeling with that set of
customers there will also be comments about UNIX.
... snip ... top of post, old email index
past posts in this thread:
https://www.garlic.com/~lynn/2011b.html#19 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#20 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#23 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#29 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#33 A brief history of CMS/XA, part 1
https://www.garlic.com/~lynn/2011b.html#47 A brief history of CMS/XA, part 1
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 04 Feb 2011 Subject: Productivity And Bubbles Blog: IBMersre:
The initial draft of Basel-II had new "qualitative section" ... basically in addition to old standard quantitative measures for risk adjusted capital ... "qualitative" would have required something slightly analogous to iso9000 ... top financial executives demonstrate that they understood the end-to-end business processes. during the review process ... it was all effectively eliminated. There are lots of complaints about the level of risk adjusted capital in Basel3 ... but somebody observed recently that most european institutions already are operating with nearly twice the risk adjusted capital called for in Basel3 (while comments that if too-big-to-fail institutions had to meet basel3 risk adjusted capital it would tank the economy).
This morning on business cable station ... they had some presidential economic adviser from 30-40 yrs ago ... who claimed that there has been lots of number fiddling going on to obfuscate and misdirect what has happened. One comment was that wallstreet and market has had all sorts of stimulus and help ... but mainstreet hasn't ... eliminating all the statistics and hype ... supposedly the total number employed at the moment is the same as were employed in the 90s (factor out all the ups & downs, the economy is stagnant for the last 15yrs or so (while population has increased significantly).
There have been comments that this has been substitute for doing (30s)
Pecora hearings and having to take any substantive action ... along
with jokes about the (tens?) billions spent on FIRE ... financial,
insurance, real-estate ... lobbying over the past 15yrs.
https://www.amazon.com/Financial-Crisis-Inquiry-Report-ebook/dp/B004KZP18M
When everything was falling apart here in the US ... there were references to how far around the world would the effects propagate. This was about the same time there was article about ratio of US executive compensation to worker compensation had exploded to 400:1 after having been 20:1 for a long time and 10:1 in most of the rest of the world. An article was written claiming that the too-big-to-fail problems wouldn't hit china because their financial institutions were much better regulated AND also that their financial executives had total compensation in the hundred of thousand range (not tens & hundreds of millions). There was also some claim that concessions/benefits that wallstreet received from congress was over 1000 times what they spent "lobbying".
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 04 Feb 2011 Subject: Credit cards with a proximity wifi chip can be as safe as walking around with your credit card number on a poster Blog: LinkedInre:
the recent, re-written magstripe wiki still retains some los gatos
lab reference
https://en.wikipedia.org/wiki/Magnetic_stripe_card
IBM has a new "100 yr" history site ... and has magstripe entry
http://www.ibm.com/ibm100/us/en/icons/magnetic/
but doesn't mention los gatos lab involvement.
the ATM 3624 wiki still mentions los gatos lab:
https://en.wikipedia.org/wiki/IBM_3624
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: Speed of Old Hard Disks Newsgroups: alt.folklore.computers Date: Fri, 04 Feb 2011 18:18:28 -0500re:
more topic drift ... mostly NSF backbone stuff ... but also references
getting suck into doing this stuff I was also working for large
numbers of processors in clusters ... the cluster stuff gets somewhat
stalled but is revived in later ha/cmp
https://www.garlic.com/~lynn/subtopic.html#hacmp
of course all of this is before internal politics shutting most of it
down (at least externally). Then head of NSF tried to help with letter
to corporation ... but that just aggravated the internal politics; ref
https://www.garlic.com/~lynn/2006s.html#email860417
old nsfnet email
https://www.garlic.com/~lynn/lhwemail.html#nsfnet
As previously mentioned, I couldn't do the HSDT pitch at NSF because
I was also doing all this stuff for processor clusters involving
lots of meetings on the east coast; referenced in the attached.
Date: 03/25/85 09:05:58
From: wheeler
To: UCB marketing rep
re: hsdt; my phone is still not being answered so I'm not getting any
messages. I was on east coast tues-friday of last week. Something has
come up & will have the same schedule this week. Will also be back
there the week of april 19th.
HSDT pitch to NSF & Erich Bloch last week went extremely well. He not
only wants to have a back-bone system to tie together all the
super-computer centers ... but NSF appeared to have been extremely
impressed with the technology and would like to work with us at a
technology level.
How is the UCB scheduling going???
... snip ... top of post, old email index, NSFNET email
Date: 03/25/85 16:11:50
From: wheeler
To: other UCB marketing rep
I'm currently available on the 15th or anytime the week of the 22-26.
Presentation will essentially be same given to head of NSF for use as
backbone to tie all supercompter centers together. Will also be making
the proposal for a couple of university type applications in addition
to using it as backbone for govenment labs. (i.e. livermore,
etc.). Block also expressed interest in the research aspects of the
technology we are using for communicating.
... snip ... top of post, old email index, NSFNET email
Date: 03/26/85 16:00:23
From: UCB marketing rep
To: wheeler
Cc: other UCB marketing rep
Lynn
I received you note of 3-25-85 re. your presentations to NSF. We have
generated interest in the Vice Chancelorfor Computing, Raymond Neff.
Could you provide us with a couple of dates the week of April 12 for
planning purposes.
... snip ... top of post, old email index, NSFNET email
Date: 04/02/85 11:02:50
From: wheeler
HSDT was pitched to head of NSF & some of the staff about a week and a
half ago. There were very favorably impressed ... NSF would like to use
it to tie together all the super computer centers (backbone
network). They also expressed a lot of interest in the HSDT technology &
would like to have some meetings out here in SJR with various NSF
technical people ... if possible have some joint projects.
I'm meeting with some ACIS people this afternoon who have heard of
the results and are anxious to use HSDT as a lever to get IBM more
involved with the super computer center.
I'm also presenting HSDT to the UofC people next thursday in Berkeley.
UofC has interest in using HSDT to tie together their campuses and
also use it as a "feeder" to the San Diego super computer center.
... snip ... top of post, old email index, NSFNET email
other old hsdt email
https://www.garlic.com/~lynn/lhwemail.html#hsdt
and later cluster scale-up became the supercomputer
https://www.garlic.com/~lynn/lhwemail.html#medusa
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 06 Feb, 2011 Subject: Productivity And Bubbles Blog: IBMersre:
Seeing the review process for Basel-2 cut a lot of stuff out from the initial draft ... and then various regulators (especially in the US) exempt lots of stuff from Basel consideration (by allowing it to be carried off book/balance) ... I can't really blame BIS. Some observation about European institutions carrying higher risk adjusted capital than even called for in Basel-3 ... while lots in the US claiming Basel-3 would tank the economy ... would tend to highlight opposing national country forces, that BIS has to deal with.
In that sense US regulators would be in collusion with the too-big-to-fail ... institutions permitting the four largest too-big-to-fail to have $5.2T in toxic assets being carried off balance at the end of 2008. Note that some number of transactions involving tens of billions of those toxic assets had sold earlier in 2008 at 22cents on the dollar. If those institutions had been required to bring those toxic assets back "on book" ... they would have been declared insolvent and had to be liquidated (way beyond anything they would have been capable of by allocating additional risk adjusted capital).
As mentioned earlier ... it wasn't like the issues weren't well
understood ... as per long-winded post from jan1999
https://www.garlic.com/~lynn/aepay3.htm#riskm
and some more old GEEK ... leading up to NSFNET backbone (and modern
internet)
http://lnkd.in/JRVnGk
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Sun, 06 Feb 2011 10:40:58 -0500"Joe Morris" <j.c.morris@verizon.net> writes:
then you get into 91, 95, & 195.
I never used 360/50 and don't have any recollection of any statement about emulation features on 50. I do know that there was special microcode assist for CPS (conversational programming system, ran under os/360). Also, science center had wanted a 360/50 to build prototype virtual memory (pending availability of 360/67) but had to settle for 360/40 (doing cp/40 before morphing to cp/67 when 360/67 became available) because all the available 360/50s were going to FAA ATC. There would be statement in 360/50 functional characteristic ... but there isn't one up at bitsaver.
there is some folklore that the big cutover from lease to purchase in the early 70s was some executive about ready to retire and wanted to really boost their bonus as he was going out the door.
long ago and far away I was told that in the gov. hearings into ibm ... somebody from RCA(?) testified that all the computer vendors realized by the late 50s that the single most important feature was to have a single compatible architecture across the complete machine line (businesses were in period of large growth ... start with smaller machine and then having to move up to larger machines ... really big inhibitor to market was software application development, being able to re-use applications significantly helped selling larger machines). The statement wase that IBM had the only upper management that were able to force the individual plant managers (responsible for different product lines) to conform to common architecture. While IBM may have lost out in some competition for specific machines in niche markets ... being the only vendor selling the single, most important feature ... common architecture ... allowed IBM to dominate the market (only vendor selling the single, most important feature ... would have allowed them to get nearly everything else wrong and still be able to dominate the competition).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Other early NSFNET backbone Newsgroups: alt.folklore.computers Date: Sun, 06 Feb 2011 11:46:54 -0500re:
and
Date: 11/06/85 11:22:22
From: wheeler
To: ibm earn
Is there any more follow-up about IBM/Europe supporting a SLAC/CERN
link?? The HSDT backbone for the super computer centers is
progressing. Currently ACIS is "carrying" the proposal for an NSF
joint study with HSDT nodes at each of the centers. In addition, I'm
going up to Berkeley on Friday to discuss a bay area node(s) for
access to the super computer centers. Berkeley & Stanford are trying
to co-ordinate something. It might be possible for there to be a
tie-in between SLAC and the Stanford legs. That could make the CERN
connection easier.
... snip ... top of post, old email index, NSFNET email
The above ibm earn email was to person that had sent this email
the year before:
https://www.garlic.com/~lynn/2001h.html#email840320
As previously mentioned ... internal politics resulting in all sort of
road blocks and stalling and eventually shutdown external activity.
various NSFNET related email
https://www.garlic.com/~lynn/lhwemail.html#nsfnet
Note, SLAC/CERN were "sister" institutions, sharing some amount of
software and staff tending to visit each other (&/or take sabbaticals
at each others institutions) frequently. SLACs reference to 1st
website outside CERN:
http://www.slac.stanford.edu/history/earlyweb/index.htm
SLAC was big vm370 institution and for a long time hosted BAYBUNCH
(monthly bay area vm370 user group meeting) ... SLAC was on Sand Hill
road and had linear accelerator that crossed beneath interstate 680
(aka Stanford Linear Accelerator Center)
http://www.slac.stanford.edu/
Berkeley 10m telescope was renamed "Keck" when they got funding from
Keck foundation
http://www.keckobservatory.org/
https://en.wikipedia.org/wiki/W._M._Keck_Observatory
EARN was the European equivalent to BITNET ... some past posts
https://www.garlic.com/~lynn/subnetwork.html#bitnet
misc. past posts mentioning HSDT
https://www.garlic.com/~lynn/subnetwork.html#hsdt
NCAR also had one of the early NAS/SAN using HYPERchannel and remote
device adapters. IBM mainframe staged data to IBM disks ... and then
loaded disk "channel programs" into HYPERchannel (A515) remote device
adapter (simulated IBM mainframe channel) which various other
(non-ibm) mainframes in the HYPERchannel network could invoke
(allowing direct data transfer from the ibm disks to the non-ibm
mainframes w/o having to pass through the ibm mainframe).
http://www.ucar.edu/educ_outreach/visit/
In the early 90s with prompting ... lots of gov. labs would attempt to commercialize internal technologies ... with the NCAR NAS/SAN being ported to AIX and being marketed as "Mesa Archival".
NCAR's NAS/SAN was also largely behind the latter work to support "3rd party transfers" for HiPPI switches & IPI disks.
For other drift Los Alamos had technology that was packaged and being
marketed by General Atomics (in san diego). General Atomics eventually
also had contract to operate the UCSD supercomputer center. When we
were working with Livermore commercializing their filesystem as
Unitree (in conjunction with HA/CMP) ... there was a lot of activity
with General Atomics.
http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?reload=true&arnumber=113582
more recent
http://www.datastorageconnection.com/article.mvc/UniTree-A-Closer-Look-At-Solving-The-Data-Sto-0001
misc. past posts mentioning "Mesa Archival":
https://www.garlic.com/~lynn/2001.html#21 Disk caching and file systems. Disk history...people forget
https://www.garlic.com/~lynn/2001.html#22 Disk caching and file systems. Disk history...people forget
https://www.garlic.com/~lynn/2001f.html#66 commodity storage servers
https://www.garlic.com/~lynn/2002e.html#46 What goes into a 3090?
https://www.garlic.com/~lynn/2002g.html#61 GE 625/635 Reference + Smart Hardware
https://www.garlic.com/~lynn/2003b.html#29 360/370 disk drives
https://www.garlic.com/~lynn/2003b.html#31 360/370 disk drives
https://www.garlic.com/~lynn/2003h.html#6 IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
https://www.garlic.com/~lynn/2004d.html#75 DASD Architecture of the future
https://www.garlic.com/~lynn/2004p.html#29 FW: Is FICON good enough, or is it the only choice we get?
https://www.garlic.com/~lynn/2005e.html#12 Device and channel
https://www.garlic.com/~lynn/2005e.html#15 Device and channel
https://www.garlic.com/~lynn/2005e.html#16 Device and channel
https://www.garlic.com/~lynn/2005e.html#19 Device and channel
https://www.garlic.com/~lynn/2006n.html#29 CRAM, DataCell, and 3850
https://www.garlic.com/~lynn/2007j.html#47 IBM Unionization
https://www.garlic.com/~lynn/2008p.html#51 Barbless
https://www.garlic.com/~lynn/2009k.html#58 Disksize history question
https://www.garlic.com/~lynn/2009s.html#42 Larrabee delayed: anyone know what's happening?
https://www.garlic.com/~lynn/2010d.html#69 LPARs: More or Less?
https://www.garlic.com/~lynn/2010d.html#71 LPARs: More or Less?
https://www.garlic.com/~lynn/2010m.html#85 3270 Emulator Software
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 06 Feb, 2011 Subject: Productivity And Bubbles Blog: IBMersre:
Asset management division's stumbles tarnish Goldman Sachs
http://www.washingtonpost.com/wp-dyn/content/article/2011/02/05/AR2011020504529.html
Paper version of the above had comment that it is all a game to them ... money is just how score is kept. One of the "game" metaphors during the height of the bubble was "musical chairs" ... wondering who would be left holding the toxic assets when the bubble burst (music stopped).
Goldman's TARP funds came about the time and about the same as 2008
compensation payout for having lost money in 2008:
http://abcnews.go.com/WN/Business/story?id=6498680&page=1
from above:
Goldman Sachs, which accepted $10 billion in government money, and
lost $2.1 billion last quarter, announced Tuesday that it handed out
$10.93 billion in benefits, bonuses, and compensation for the year.
... snip ...
Goldman was also one of the investment banks that got bank charters
(in theory would have been precluded under the "prime purpose" of GLBA,
discussed up thread). Easy federal reserve funds helped them (and
others) make a bundle and to pay back TARP funds. Could be considered
trying to help preserve the 400+% spike in wallstreet bonuses during
the bubble (even after the bubble had collapsed and institutions were
loosing money).
http://www.businessweek.com/#missing-article
Note that the original regulator that was directed by Reagan to cut
the reserve requirements in half refused ... and was replaced by
somebody that would follow directions. Much of the freed up reserves
then disappeared into wallstreet, never to be seen again (and was
major part of S&L crisis). That regulator then was rewarded with
very lucrative prestigious job on wallstreet. some of this also
discussed in jan1999 post
https://www.garlic.com/~lynn/aepay3.htm#riskm
Note in longwinded thread in (linkedin) "Financial Drime Risk, Fraud
and Security" group (still closed) titled "What do you think about
fraud prevention in the governments?" ... there is mention about a
recent book "America's Defense Meltdown" ... but the venality of
wallstreet dwarfs the pentagon's ("13 Bankers: The Wallstreet Takeover
and the Next Financial Meltdown" and "Griftopia; Bubble Machines,
Vampire Squids, and the Long Con That is Breaking America"). Piece of
it is archived in this post:
https://www.garlic.com/~lynn/2011.html#53
The "Defense Meltdown" frequently cites Boyd ... I had sponsored
Boyd's briefings at IBM ... and some of the pentagon/wallstreet is
also mentioned in (open linkedin) Boyd discussion group
http://www.linkedin.com/groups/Disciples-Boyds-Strategy-1015727?trk=myg_ugrp_ovr
"America's Defense Meltdown" ... open URL
http://lnkd.in/XtjMSM
Especially the "Griftopia" reference goes into some detail that the parasitic/game activity has been going on forever ... periodically getting out of control and nearly killing the host (normally parasites are dependent on their hosts, killing the host can result in their own demise).
There have been some published studies that claims a large percentage are amoral (going well beyond simple greedy) ... which would require large amounts of adult supervision.
the other periodic reference is sociopath and/or amoral sociopath ... the issue is they know how they got their money ... but claim is that many have no sense that ethics or morality is any way involved. another study claimed that small percent of children make it past 4-5 yrs old w/o acquiring a sense of morality ... and if they don't have it by then ... they grow up to never have it.
slightly related ... but different ... are periodic references to business ethics is an oxymoron ... where people that appear to otherwise have a sense of morality ... don't see it as applicable in a business context.
in any case, there can be no expectation of do right by these individuals (requiring large amount of structured environment and oversight).
quote recently seen on facebook
FDR in 1936:"We know now that Government by organized money is just as dangerous as Government by organized mob." Obama in 2011:????
for a little more background, two trillion dollar meltdown starts
about 1970 and moves forward ... looking at all sorts of happenings
over the past 40 yrs (kindle version on sale only $2.88)
https://www.amazon.com/Two-Trillion-Dollar-Meltdown-Rollers-ebook/dp/B0097DE7DM/
Does mention regulator under Reagan was major factor in S&L crisis ... but deflects some attention from Reagan ... not mentioning replaced the previous regulator who wouldn't do what his president asked.
some more (old geek) discussion on early NSFNET backbone (operational
precursor to modern internet) ... but in a.f.c. newsgroup:
https://www.garlic.com/~lynn/2011b.html#58
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: A Two Way Non-repudiation Contract Exchange Scheme. Newsgroups: sci.crypt Date: Sun, 06 Feb 2011 15:44:11 -0500we were called in to help wordsmith the cal. electronic signature legislation. there seemed to be a lot of lobbying by vendors related to digital signature technology to obfuscate the issues ... possibly hoping to confuse the issue because "human signature" and "digital signature" both contained the word "signature".
the lawyers working on the legislation, very clearly spelled out that non-repudiation is a legal issue, not a technology issue ... and that "human signature" requires some indication of intent, that a human has read, understood, agrees, approves, and/or authorizes (related to appearance of "human signature" as part of legally resolving a dispute).
in terminal point-of-sale ... it has been pointed out that entering the PIN is a form of authentication ... but the screen that asks if you approve the transaction ... and you have to click yes or no ... the "clicking" yes ... is the equivalent of the "human signature" (independent of swiping card & entering the PIN as form of authentcation).
in the 90s, there was some attempt to reduce resistance to deploying "digital signature" technology for payment transactions ... by claiming that digitally signed (aka authenticated) transactions could be considered "non-repudiaton" and therefor could be used to bypass REG-E ... and change the burden of proof in a dispute from the merchant to consumer. however, that still appeared to leave the consumer with the privileged of paying $100/annum for digital signature technology that, if used, would mean that in any dispute ... the burden of proof would be shifted from the merchant to consumer (the consumer looses out all around ... and for the merchant it would be a huge savings, which then would cost justify the merchant deploying the necessary digital signature technology).
trying to get legislation to mandate use of "digital signatures" as the *ONLY* "electronic signature" was part of overcoming consumer resistance to the $100/annum expense (along with obfuscating the issue that "digital signature" should be considered equivalent to "human signature") ... but that still left the issue why a consumer would want to accept the burden of proof in dispute (involving payment transaction).
re:
https://www.garlic.com/~lynn/subpubkey.html#signature
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 06 Feb, 2011 Subject: VM13025 ... zombie/hung users Blog: z/VMThe summer I was at Boeing, I had done pageable kernel support for cp67. While it never shipped in cp67 product (lots of other stuff I had done as undergraduate &/or before joining IBM, pageable kernel support made it out for VM370 ... but never for CP67). recent reference to Boeing:
However, I did include pageable kernel support in the internal production CP67 (after joining the science center). Part of the pageable kernel support was creating a "SYSTEM" virtual memory table (for the kernel map of fixed & pageable components) and a "SYSTEM" utable (aka VMBLOK) to go with it.
Later, one of the things Charlie (invented compare&swap instruction
while working on fine-grain multiprocessor locking, CAS was chosen
because they are Charlie's initials) did was rewrite kernel virtual
machine serialization (eliminating lots of zombie users as well as
crashes from dangling activity after the virtual machine was
gone). Part of the serialization rewrite was to re-assign pending
operations to the "SYSTEM" utable (allowing user to complete reset,
eliminating lots of zombies w/o having to worry about system crashes
from outstanding activity waiting to complete). misc. past posts
mentioning SMP and/or compare&swap instruction
https://www.garlic.com/~lynn/subtopic.html#smp
As part of converting from cp67 to vm370 ... mentioned in this old email:
https://www.garlic.com/~lynn/2006v.html#email731212 ,
https://www.garlic.com/~lynn/2006w.html#email750102 ,
https://www.garlic.com/~lynn/2006w.html#email750430
some other work on automated benchmarking was also moved ... basically synthetic workload that could be parameterised to match a whole lot of different characteristics. One of the things was "extremely heavy" ... which was possibly ten times the load normally seen in real life ... which would always result in vm370 system crashing.
So as part of eliminating all possible crashes (in the cp67 to vm370
migration) ... included rewriting virtual machine serialization
... using the technique Charlie had used for cp67. Eventually all
known cases of system crashes had been eliminated and the final set of
automated benchmarks that was part of releasing my Resource Manager
(vm370 release 3 plc 9) involved 2000 benchmarks that took three
months elapsed time to run. misc. past posts mentioning automated
benchmarking
https://www.garlic.com/~lynn/submain.html#benchmark
While 23Jun69 unbundling had started charging for application
software, but they made the case that kernel software was still
free. In the wake of FS demise (and considered responsible for
allowing clone processors to get market foothold), misc. past
mentioning FS
https://www.garlic.com/~lynn/submain.html#futuresys
There was then decision to start charging for kernel software ... and
my Resource Manager was selected to be the guinea pig. misc. past
posts mentioning resource manager
https://www.garlic.com/~lynn/subtopic.html#fairshare
misc. past posts mentioning 23Jun69 unbundling
https://www.garlic.com/~lynn/submain.html#unbundle
For vm370 release 4, it was decided to release multiprocessor support and a design was chosen based on some efforts that weren't released that I was involved with dependent on lots of stuff in the Resource Manager. The transition guidelines for kernel software charging allowed direct hardware support to still be free (aka multiprocessor support) and it couldn't have dependency on charged for software (aka my Resource Manager). The eventual solution was to take approx. ninety percent of the lines of code in my Resource Manager and move it into the free release 4 base (leaving the release 4 Resource Manager the same price as the release 3 Resource manager, even though only ten percent of the code).
So I believe that for extended period of time (starting with Resource
Manager) all cases of zombie users and crashes because of dangling
activity (after user was logged off) had been eliminated ... until
some APAR to DMKDSP that was VM13025. This APAR re-introduced zombie
users by allowing some delay in resetting a virtual machine
... because of poorly thought out analysis involving some system
failure.
Date: 02/17/86 09:49:56
From: wheeler
To: melinda
re: hunguser; I've just shot a couple of such conditions via the
internal IBMVM forum with people appending DUMPRX output. You might
check 13025 test in DMKDSP which appears to be contributing to most of
the problems (I haven't tracked it back with change team why test was
put in ... I presume it was some sort of unpredictable abend with
logoff finishing while something was active for user).
One specific case causing hunguser was virtual machine in TIO-LOOP
condition on virtual console ... some sort of error caused DMKGRF to
clear busy condition then exiting w/o turning off TIO-LOOP flag and
VMEXWAIT. It is likely that there may be a couple other similar
situation around CP that weren't causing hangs prior to 13025 (note:
CFP/CFQ clears VMEXWAIT, if user does system-reset#logoff ... 13025
won't catch him).
Completion of logoff with outstanding activity should be caught by
CFP/CFQ and/or PGSPW (PTRPW). Situations where it won't work are with
routines that do SWTCHVMs and have pending activity for another
VMBLOK. Do you remember Jerry S. description of delayed spool
interrupt for batch machines. I worked on a fix with him for the
problem. That situation could also result in delayed activity for a
logged-off VMBLOK, ... there is nothing obvious that CFP/CFQ/PGSPW can
catch while the "sending" VMBLOK is waiting on the completion of the
spooled console page I/O for the "receiving" VMBLOK. There are likely
similar situations in VMCF/IUCV processing.
... snip ... top of post, old email index
Date: 02/17/86 11:00:21
From: wheeler
To: melinda
re: 13025; oh yes, there is also a 13025 test in uso ... where if
VMEXWAIT is on, it simulates the FORCE command instead off LOGOFF ...
... snip ... top of post, old email index
The reference in the above to DUMPRX was a replacement for IPCS that I
had done in REXX ... some past posts
https://www.garlic.com/~lynn/submain.html#dumprx
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 06 Feb, 2011 Subject: vm/370 3081 Blog: IBM Historic Computingre:
diag sio
4341 9.52 15.28
3081 3.45 7.18
both machines were running hpo so part of 4341-2 ecps is "crippled" or
4341 numbers should have been better (3081 has the latest cpu speed-up
EC on it also).
old post mentioning original work for endicott/4341 microcode (ECPS):
https://www.garlic.com/~lynn/94.html#21 370 ECPS VM microcode assist
part of the above (aka increasing bloat in diag i/o) contributes to my
paged mapped cms filesystem pathlengths looking better. however, there
is the fundamental philosiphical architecutre issue ... even at
optimal, "diag i/o" is still using (real address) channel program
paradigm which requires overhead to remap into a virtual
paradigm. paged mapping already takes care of all that before it ever
hits the cp interface. misc. past post mentioning having originally
done the work on cp67/cms and then ported to vm370
https://www.garlic.com/~lynn/submain.html#mmap
old post mentioning having done the precursor to diag i/o while
undergraduate:
https://www.garlic.com/~lynn/99.html#95 Early interupts on mainframes
https://www.garlic.com/~lynn/2000c.html#42 Domainatrix - the final word
https://www.garlic.com/~lynn/2003.html#60 MIDAS
https://www.garlic.com/~lynn/2003b.html#33 dasd full cylinder transfer (long post warning)
https://www.garlic.com/~lynn/2003k.html#7 What is timesharing, anyway?
https://www.garlic.com/~lynn/2004d.html#66 System/360 40 years old today
https://www.garlic.com/~lynn/2005j.html#54 Q ALLOC PAGE vs. CP Q ALLOC vs ESAMAP
https://www.garlic.com/~lynn/2005t.html#8 2nd level install - duplicate volsers
https://www.garlic.com/~lynn/2008r.html#33 What if the computers went back to the '70s too?
https://www.garlic.com/~lynn/2008s.html#56 Computer History Museum
https://www.garlic.com/~lynn/2010l.html#29 zPDT paper
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Mon, 07 Feb 2011 09:31:44 -0500Peter Flass <Peter_Flass@Yahoo.com> writes:
by at least the mid-70s, the resource trade-off had inverted and FBA
devices were being produced produced ... however the POK (batch)
favorite son operating system didn't/wouldn't support them. I was told
that even giving them fully tested and integrated FBA support, I would
still have to come up with a $26M business case to cover the costs of
documentation and education. I wasn't allowed to use life-cycle savings
or other items ... but had to be purely based on incremental new disk
sales (new gross on the order of $200m-$300m new disk sales). The
argument was then that existing customers would just buy the same amount
of FBA as they were buying CKD. misc. past posts mentioning FBA, CKD,
multi-track search, etc
https://www.garlic.com/~lynn/submain.html#dasd
There was a similar/analogous shift in real resources with uptake in
RDBMS in the 80s. In the 70s, there was some skirmishes between the
'60s IMS physical database group in STL and the RDBMS System/R group
in bldg. 28 ... some old posts mentioning original sql/relational
https://www.garlic.com/~lynn/submain.html#systemr
IMS group claiming RDBMS doubled the physical disk requirements for the implicit, on-disk index, index processing also significantly increasing the number of disk i/os. Relational retort was that exposing explicit record pointers in application significantly increased administrative overhead and human effort in manual maintenance. In the 80s, significant drop in disk price/mbyte minimized the additional disk space argument, and further increases in available of real storage allowed caching of index structure ... mitigating the number of physical disk I/Os (while human experience/skills to manage IMS were becoming scarce/expensive).
The early 80s also saw big explosion in mid-range business ... 43xx machines (also saw similar explosion in vax/vms). The high-end disk was CKD 3380 ... but the only mid-range disk was FBA 3370. This made it difficult for the POK favorite son operating system to play in the mid-range with 4341 ... since there wasn't a mid-range CKD product (there was some POK favorite son operating system on 4341 for some installations that upgraded from older 370s to 4341 and retaining existing legacy CKD). Eventually attempting to address the exploding mid-range opportunity (for POK favorite son operating system), the CKD 3375 was produced, which was CKD simulation on FBA 3370.
Of course, today ... all CKD devices are simulated on FBA disks.
misc. recent posts mentioning the $26M business case
https://www.garlic.com/~lynn/2011.html#23 zLinux OR Linux on zEnterprise Blade Extension???
https://www.garlic.com/~lynn/2011.html#35 CKD DASD
https://www.garlic.com/~lynn/2011b.html#47 A brief history of CMS/XA, part 1
recent old email item mentioning 4341 & 3081
https://www.garlic.com/~lynn/2011b.html#email841012
in this (linkedin) discussion about 3081
https://www.garlic.com/~lynn/2011b.html#49 vm/370 3081
https://www.garlic.com/~lynn/2011b.html#62 vm/370 3081
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Mon, 07 Feb 2011 12:30:41 -0500despen writes:
3278/3274 (DFT) move a lot of the electronics back into the controller ... which resulted in significantly reducing the transmission efficiency.
This showed up in things like signicant response time differences
between ANR and DFT (if you worried about such things, when we tried
to escalate ... the response was 3278/3274 was NOT designed for
interactive computing ... but for data-entry ... i.e. essentially
computerized keypunch). An old comparison
https://www.garlic.com/~lynn/2001m.html#19 3270 protocol
It later shows up in upload/download throughput with PC terminal emulation ... comparing thruput with a 3277/ANR emulation card vis-a-vis 3278/DFT emulation card.
All 327x was designed with star-wired coax ... with individual coax
cable from the datacenter out to every 327x terminal. Some buildings
were running into loading limits because of the enormous aggregate
weight of all those 327x cables. Token-ring was somewhat then positioned
to address the enormous problem with the weight of 327x cables. Run CAT4
out to local MAU box in departmental closet ... then star-wired CAT4
from departmental MAU box to individual terminals. Communication
division was having token-ring cards (first 4mbit/sec and later
16mbit/sec) designed with throughput oriented towards "terminal
emulation" paradigm and hundreds of such stations sharing common
LAN bandwidth. misc. past posts mentioning "terminal emulation"
https://www.garlic.com/~lynn/subnetwork.html#emulation
above posts also periodically mentioning that the hard stance on preserving the "terminal emulation" paradigm (and install base) was starting to isolate the datacenter from the growing distributed computing activity. In the late 80s, a senior disk engineer got a talk scheduled at the internal, world-wide, annual communication group conference ... and opened the talk with the statement that the communication group was going to be responsible for the demise of the disk division. While the "terminal emulation" paradigm help with early uptake of PCs (business could get a PC with 3270 emulation for about the same price as a real 3270 ... and in the same desktop footprint get both business terminal and some local computing, since the 3270 terminals were already business justified, it was no brainer business justification to switch from real teraminal to PC), the rigid stance on preserving "terminal emulation" was started to result in large amounts of data leaking out of the datacenter to more "distributed computing" friendly platforms (the leading edge of this wave was showing up in decline in disk sales).
Because the RS6000 had microchannel ... they were being forced into using cards designed for PS2 market. The workstation group had done their own 4mbit token-ring card for the PC/RT (16bit AT bus). It turns out that 16mbit PS2 microchannel card (designed for terminal emulation market) had lower per card thruput than the PC/RT 4mbit token ring card. RS6000 had similar issues with the PS2 scsi controller cards and the PS2 display adapter cards (none of them designed for high-performance workstation market).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Mon, 07 Feb 2011 12:49:12 -0500Roland Hutchinson <my.spamtrap@verizon.net> writes:
also included in this linkedin reference:
http://lnkd.in/JRVnGk
similar motivation to the group attempting to preserve the "terminal
emulation" paradigm mentioned in this post:
https://www.garlic.com/~lynn/2011b.html#64 If IBM Hadn't Bet the Company
and these
https://www.garlic.com/~lynn/subnetwork.html#emulation
also shows up as part of effort to get the corporate internal network
converted to SNA/VTAM ... part of the justification telling the
executive committee that PROFS was a VTAM application:
https://www.garlic.com/~lynn/2006x.html#email870302
in this post
https://www.garlic.com/~lynn/2006w.html#7 vmshare
and another old email
https://www.garlic.com/~lynn/2011.html#email870306
in this post
https://www.garlic.com/~lynn/2011.html#4 Is email dead? What do you think?
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 07 Feb, 2011 Subject: Boeing Plant 2 ... End of an Era Blog: Old Geek Registryre:
The observation about "cloud computing" has been repeated a number of
times regarding the various (virtual machine based) timesharing
services that sprung up in the 60s (not just BCS but also NCSS & IDC
and later TYMSHARE). In addition, there were loads of "in-house"
operations ... including the one referenced here:
https://web.archive.org/web/20090117083033/http://www.nsa.gov/research/selinux/list-archive/0409/8362.shtml
One of the challenges in the 60s for (virtual machine based) commercial timesharing being up 7x24 was the CPU meter. Machines were leased and monthly charges were based on the CPU meter (in blocks of 1st shift, 1+2 shift, 1+2+3 shift & all four shifts, aka 7x/24) which ran whenever the CPU was executing and/or there was active I/O. One of the tricks was to come up with an I/O sequence that would accept incoming characters ... but allow the CPU meter to stop when things were otherwise idle. Another challenge was to support darkroom operation ... being able to leave system up offshift w/o requiring onsite operator.
Early days there was sparce offshift use ... but it was chicken&egg; little offshift used was hard to justify being up 7x/24 ... but w/o 7x24 availability, it was difficult to encourage offshift use (so it was important early on to minimize the offshift operating costs).
past posts mentioning early commercial time-sharing
https://www.garlic.com/~lynn/submain.html#timeshare
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Mon, 07 Feb 2011 15:28:50 -0500despen writes:
A class of (3270) "psuedo devices" were added to vm370 ... and along with PVM (passthrough virtual machine) ... allowed remote terminal emulation ("dial PVM" and then select another network node to "logon" to.
Fairly early ... the person responsible for internal email client (VMSG, which was also used as the embedded technology for PROFS email) wrote PARASITE & STORY ... small compact programmable interface to PVM along with a HLLAPI-like language (predating IBM/PC and 3270 terminal emulation).
old posts with some old PARASITE/STORY information ... along
with STORY semantics and example scripts:
https://www.garlic.com/~lynn/2001k.html#35 Newbie TOPS-10 7.03 question
https://www.garlic.com/~lynn/2001k.html#36 Newbie TOPS-10 7.03 question
including things like logging into the Field Engineering "RETAIN" system and pulling off lists of service support information.
both PARASITE & STORY are remarkable for the function they provided ... considering the size of their implementation.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 07 Feb, 2011 Subject: vm/370 3081 Blog: IBM Historic Computingre:
I have no knowledge of VM having such problems (modulo any special
case code introduced in vm/sp1 for the ACP/TPF case) ... this is long
winded post to (linkedin) z/VM that includes some discussion of
original release 4 multiprocessor support. The precursor had been for
a 5-way implementation ... and then a 16-way implementation (neither
announced). The 16-way had been going great guns ... including
co-opt'ing some of the 3033 processor engineers spending spare time on
the project. Then somebody informed the head of POK that it might take
decades before the POK favorite son operating system had 16-way
support ... at which point the head of POK invited some people to
never visit POK again ... and that the 3033 processor engineers should
focus solely on what they were doing.
https://www.garlic.com/~lynn/2011b.html#61
misc. past posts specifically mentioning 5-way effort
https://www.garlic.com/~lynn/submain.html#bounce
general posts mentioning multiprocessor
(&/or compare&swap instruction)
https://www.garlic.com/~lynn/subtopic.html#smp
The 370 strong memory consistency takes significant toll on cache processing. Both standard MVS & VM had an issue with kernel dynamic storage where it was possible for different allocated storage could share a common cache line. If the two different storage areas were in use by different processors concurrently ... there could be significant cache-line thrashing between the different processors. This issue existed on 2-way 3081 ... but probability of it happening increased for 4-way 3084. To address the problem, both MVS and VM had their kernel dynamic storage modified to be cache line sensitive (storage allocation started on cache-line boundary and was allocated in multiples of cache line). This change (kernel dynamic storage) got something like 5-6% performance improvement.
Past post mentioning MVS smp getting extended from 2-way to 4-way (for
3084) as well as well as cache-line sensitivity
https://www.garlic.com/~lynn/2001j.html#18
https://www.garlic.com/~lynn/2003j.html#42
https://www.garlic.com/~lynn/2007.html#44
https://www.garlic.com/~lynn/2008e.html#40
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 07 Feb, 2011 Subject: Boeing Plant 2 ... End of an Era Blog: Old Geek Registryre:
One of the biggest users of cmsapl (from cambridge science center)
... and then aplcms (from palo alto science center) .. and then
various other generations of APL ... was the (virtual machine) online
HONE system. ... the US datacenters were consolidated in bldg across
the back parking lot from PASC in the mid-70s (and in the late 70s had
one of the largest single system operations in the world). HONE
systems provide online sales & marketing support ... worldwide as HONE
systems were cloned around the world
https://www.garlic.com/~lynn/lhwemail.html#hone
past posts mentioning HONE
https://www.garlic.com/~lynn/subtopic.html#hone
random trivia ... the HONE consolidated datacenter bldg (across the back parking lot from PASC) has a different occupant now ... but may know the occupant of the new bldg that went up next door to it (1601 cal, palo alto, ca.). Some of my earliest overseas business trips was being asked to help with new HONE clones that went in at various locations around the world (including when EMEA hdqtrs moved from US to Paris).
I first met Pete (and couple others) in the early 70s when they came out to Cambridge to add 3330 & 2305 device support to "CP67I" system. The production cambridge system (360/67 CP67L) had "H" updates to provide virtual 370 virtual machines. Then (on top of CP67H) were set of "I" updates which modified CP67 to run on 370 architecture (instead of 360/67). The cp67i system was operational in 370 virtual machines a year before there was any real 370 hardware supporting virtual memory (in fact, booting the cp67i system was one of the early tests for real 370 virtual memory hardware). Internally for a long time, large number of 370 machines were running cp67i (sometimes called "cp67sj" with the added device support for 3330 & 2305).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 08 Feb, 2011 Subject: vm/370 3081 Blog: IBM Historic Computingre:
additional 3084 ... just straight storage alterations resulting in
x-cache chatter ... not even the scenario where two different storage
uses overlap in the same cache-line ... resulting in cache-line
thrashing between different machines. There is also issue that the
308x channel processor is even slower than the 303x channel director
(aka 158 microcode engine with just the 158 integrated channel
microcode and no 370 microcode).
Date: 09/17/82 10:40:29
From: wheeler
<... snip ...>
Performance numbers for the 3084 seem to have some liberties. 4-way
should have three times the performance interferance that a 2-way
(cache invalidation signals from 3 other processors instead of
one). They cheat with the 3083 versis 3081. for example, on a 158ap,
running a UP generated system ... the processor runs 10% slower if the
switch on the machine is in AP-mode rather than UP-mode (additional
delay in each machine cycle just to listen for cache invalidation
signals from the other processor ... this is w/o the other processor
even executing anything generating storage alterations & cache
invalidation signals). I've heard that the 3084 numbers are somewhat
selected benchmarks that do minimal storage alterations ... extensive
storage alteration programs can have disastrous effects on 3084
performance. ... I've been told that almost every control unit that
has attached to a 308x has had to undergo hardware ECs ... apparently
it was easier for every control unit hardware group in the company
(even on machines no longer with development group people available)
to resolve the problems than for the 308x channels. Also did you see
the message that ACP runs 20% slower on a 3081d than on a 3033. On a
3081k, ACP runs 5% faster than a 3033. POK is started a special 3081k
CPU program where the 3081s coming down the line will be tested to see
if they can run with their clock cranked down. If they pass, they will
be special high performance 3081Ks which run slightly faster than
normal 3081ks.
... snip ... top of post, old email index
aka "overclocking" by any other name.
following was extremely long discussion and has been heavily snipped.
one of the issues was 3081 "pageable" microcode having impact on
system throughput (mentioned in the SIE instruction implementation)
https://www.garlic.com/~lynn/2011.html#62 SIE - CompArch
Date: 2/10/81 14:05
To: wheeler
...
We have a new problem in cache management. The 3081 has cache lines
containing pageable microcode: about 1000 words (108 bits) of
non-pageable and about 1000 words of pageable. Unfortunately, it
appears that these 1000 words of pageable microcode are not enough to
hold both the VM and the MVS microcode for the new goodies present in
811 architecture. I'm hoping our analysis was wrong and that is
not what we are looking at, but it indeed appears to be. It may be a
few years yet before we see an 811 suitable for use by VM.
Even if we have enough microcode cache, there is a small cache size
for other work in comparison to the storage size. It was designed
with 16M in mind and has been 'upgraded' to 64M without increasing the
cache size (brilliant move). There is thought of requesting
architecture to define an instruction to cast out all altered cache
lines from the store-through cache. This would place cache management
under more direct control from the scheduler.
...
While the pageable microcode cache seems to be inadequate for both
VM/370 and VM/811 (VM/370 was hoping to catch VM/PE, but VMA+MVSA is
too tight a fit), it is much worse for VM/811 than for VM/370.
Management is still talking about 12/82 delivery dates (which they in
fact made impossible when they decided to re-convert all control
blocks, module names and linkage conventions last summer). I live in a
mad-house.
...
... snip ... top of post, old email index
... oh, with regard to "811" ... this was code name from 370/xa architecture documents carrying date of nov1978.
for all I know, there could be possibility that "vmtool" had 2-way
dependency. in the wake of the demise of Future System ... there was
mad rush to get products back into 370 product pipeline (370 work had
been pretty much been killed off during the FS period).
https://www.garlic.com/~lynn/submain.html#futuresys
Part of this was the 3033 (using 168 wiring diagram remapped to faster chips) ... in parallel with 370/xa effort. POK had managed to convince the corporation that vm370 should be killed, the 370 (burlington mall) development group shutdown, and all the people moved to POK to support MVS/XA development (otherwise MVS/XA wouldn't be able to meet its ship schedule). Part of that was an internal only virtual machine tool (never intended to be shipped to customers) supporting MVS/XA development.
Endicott managed to save the vm370 product mission, but had to reconstitute a development group from scratch.
The vmtool work went on independent of later vm370 work ... including the adding of multiprocessor support. As a result, the vmtool may have chosen a completely different SMP implementation. Later when decision was made to make limited ship of the vmtool (for a vm/xa) ... it could have required rework from a 2-way specific design to a 4-way design (in order to support 3084).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Tue, 08 Feb 2011 13:35:39 -0500Michael Wojcik <mwojcik@newsguy.com> writes:
At the time of Interop '88, for lots of people, it wasn't clear that SNMP was going to be the winner. also at Interop '88, vendors had extensive GOSIP/OSI stuff in their booths (fed. gov. was mandating internet would be killed off and be replaced by OSI stuff).
misc. past posts in this thread:
https://www.garlic.com/~lynn/2011b.html#57 If IBM Hadn't Bet the Company
https://www.garlic.com/~lynn/2011b.html#63 If IBM Hadn't Bet the Company
https://www.garlic.com/~lynn/2011b.html#64 If IBM Hadn't Bet the Company
https://www.garlic.com/~lynn/2011b.html#65 If IBM Hadn't Bet the Company
https://www.garlic.com/~lynn/2011b.html#67 If IBM Hadn't Bet the Company
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@garlic.com (Lynn Wheeler) Date: 08 Feb, 2011 Subject: IBM Future System Blog: IBM Historic Computingre:
from long ago and far way .... cp67 "G" updates ... base virtual 370
was cp67H updates, cp67 running on 370 (rather than 360/67) was cp67I
updates, 3330&2305 device support was cp67sj updates. System/R was
original sql/relational implementation
Date: 01/17/80 08:12:39
From: wheeler
There are/were three different projects.
A long time ago and far away, some people from YKT come up to CSC and
we worked with them on the 'G' updates for CP/67 (running on a 67)
which was to provide virtual 4-way 370 support. Included as part of
that was an additional 'CPU' (5-way) for debug purposes which was a
super set of the other four and could selectively request from CP that
events from the 4-way complex be sent to the super CPU for handling
(basically it was going to be a debug facility). All of this was going
to be for supporting the design & debug of a new operating
system. When FS started to get into trouble the YKT project was
canceled (something like 50 people at its peak) and the group was
moved over to work on FS in hopes that they might bail it out.
2) One or two people from that project escaped and got to SJR. Here,
primarily Vera Watson, wrote R/W shared segment support for VM (turns
out it was a simple subset of the 'G' updates) in support of System R
(relational data base language). Those updates are up and running and
somewhat distributed (a couple customers via joint studies and lots of
internal sites).
3) Independent of that work, another group at SJR has a 145 running a
modified VM with extremely enhanced modifed timer event support. The
objective is to allow somebody to specify what speed CPU and how fast
the DASD are for a virtual machine operator test. VM calculates how
fast the real CPU & DASD are and 'adjust' the relative occurance of
all events so that I/O & timer interrupts occur after the appropriate
number of instructions have been executed virtually. These changes do
not provide multiple CPU support
... snip ... top of post, old email index
past mention of cp67g
https://www.garlic.com/~lynn/2002g.html#76 Pipelining in the past
https://www.garlic.com/~lynn/2007f.html#10 Beyond multicore
recent mention of virtual 370 on cp67
https://www.garlic.com/~lynn/2011b.html#39 1971PerformanceStudies - Typical OS/MFT 40/50/65s analysed
https://www.garlic.com/~lynn/2011b.html#69 Boeing Plant 2 ... End of an Era
--
virtualization experience starting Jan1968, online at home since Mar1970
From: lynn@GARLIC.COM (Anne & Lynn Wheeler) Subject: Re: Custom programmability for 3270 emulators Newsgroups: bit.listserv.ibm-main Date: 8 Feb 2011 15:22:51 -0800charlesm@MCN.ORG (Charles Mills) writes:
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Wed, 09 Feb 2011 08:43:20 -0500Peter Flass <Peter_Flass@Yahoo.com> writes:
somewhat similar recent post in (linkedin) MainframeZone group
https://www.garlic.com/~lynn/2011.html#23 zLinux OR Linux on zEnterprise Blade Extension???
recent post in ibm 2321 (data cell) thread in (linkedin) IBM Historic
Computing group
https://www.garlic.com/~lynn/2010q.html#67
yelavich cics history pages from way back machine
https://web.archive.org/web/20050409124902/www.yelavich.com/cicshist.htm
other yelavich cics history
https://web.archive.org/web/20040705000349/http://www.yelavich.com/history/ev198001.htm
and
https://web.archive.org/web/20041023110006/http://www.yelavich.com/history/ev200402.htm
past posts mentioning CICS (&/or BDAM)
https://www.garlic.com/~lynn/submain.html#cics
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Thu, 10 Feb 2011 08:36:06 -0500hancock4 writes:
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Thu, 10 Feb 2011 09:25:20 -0500hancock4 writes:
specialized formating with keys and data ... and then search operations that would scan for specific keys &/or data ... minimizing filesystem structure in real memory. multi-track search could scan every track at same arm location (cylinder) ... but to continue would have to be restarted by software. major os/360 disk structure was VTOC (volume/disk table of contents) ... which used multi-track search. The other was library files, PDS (partitioned data set) that was used for most executable software (as well as other items) ... which had PDS "directory" ... that was also searched with multi-track search.
As part of scarce real-storage the disk i/o search argument was in real processor storage ... and was refetched by search command for every key/data encountered for the compare operation (requiring end-to-end access from disk to processor storage).
The all-time winner and complicated in this was ISAM channel programs which could do a search, read the data ... reposition the arm ... and use the previously read data for later search argument (potentially repeated several times, all in single channel program w/o interrupting the processor)
As I mentioned the trade-off started to shift by at least the mid-70s ... with real storage resources significantly increasing ... while I/O & disk thruput improvements was starting to lag significantly (becoming the major bottleneck; other infrastructures was increasingly leverage real storage to compensate for the growing disk bottleneck).
In the late 70s, I was brought into datacenter for large national retailer that had large number of loosely-coupled processors running latest os/360 in a "shared disk" environment. They were having a really horrible throughput problem and several experts from the corporation had already been brought in to examine the problem.
I was brought into classroom that had several tables covered with foot-high stacks of printed output ... all performance activity reports for all the various systems (including each system individual disk i/o counts at several minute interval). After 30-40 minutes of fanning all the reports ... I asked about one specific disk ... that under peak-load ... the aggregate sum of the disk i/o counts across all systems seemed to peg at six or seven (very low correlation of peak versus non-peak for any other item) ... was just rough measure since I was doing the aggregation across all systems and correlation in my head.
Turns out the disk contained the shared application program (PDS) library for all the regions. It had large number of applications and the size of the PDS directory was three (3330) cylinders. Every program load (across all systems) first required a multi-track search of the PDS directory ... on avg was a 1.5 cylinder search. 3330 spinning a 3600 RPM ... the first multi-track search i/o took full 19 revolutions (or 19/60th of a second) with a 2nd multi-track search of 9.5 revolutions (9.5/60 of a second) ... before getting the actual location of the PDS member application program ... movement and load maybe 2/60 of a second. In aggregate, each program load was taking 3 disk I/Os and approx. 29.5/60 of a second .... or the whole infrastructure across all processor for all the national regions for all retail stores ... was limited to loading two applications per second.
Now normally disk infrastructure has multiple disks sharing a common controller and a common channel. Because of the dependency to repeatedly reload the search argument from processor memory ... a multi-track search locks up the controller and channel for the duration of the operation (not available for any other operations) ... severely degrading all other operations.
So solution ... was replicate the program library ... one for processor (instead a shared/common for all processors) ... requiring lots for manual effort to keep them all in sync. The replicated program library was also split ... with more application logic to figure out which of the libraries contained specific application (with dedicated set of program libraries per system).
Similar, but different story about pervasive use of multi-track by os/360 (and its descendants) was san jose research for a period had a 370/168 running MVS (replacing 370/195 that had run MVT) in shared disk configuration with 370/158 running VM. Even tho, VM used CKD disks, its I/O paradigm had always been FBA (and was trivial to support real FBA disks). SJR had rule that while the disks & controller were physically shared, they were logically partitioned so I/O activity from the two different systems wouldn't interfere.
One day an operator accidentally mounted a MVS 3330 pack on a VM "string" disk controller. Within 10 minutes, operations were getting irate calls from VM users about severe degraded performance. Normal multi-track search by MVS to its pack ... was causing extended lock-out of VM access to other disks on the same string/controller). MVS operations refused to interrupt the application and move the 3330 to an MVS string/controller.
A few individuals then took a pack for a VS1 system that had been highly optimized for operation under VM ... got the pack mounted on a MVS string/controller and brought up VS1 undere VM (on 370/158) ... and started running their own multi-track searches. This managed to nearly bring the MVS system to a halt (drastically reducing MVS activity involving the 3330 on the VM string ... and significantly improve response for the VM users that were accessing data on that string). MVS operations decided that they would immediately move the MVS 3330 (instead of waiting to off-shift) ... if the VM people would halt the VS1 activity.
One of the jokes is that a large factor contributing to horrible TSO response under MVS (TSO users don't normally even realize how horrible it is ... unless they have seen VM response for comparison) ... isn't just the scheduler and other MVS operational characteristics ... but also the enormous delays imposed by multi-track search paradigm.
misc. past posts discussing CKD, multi-track search and FBA
https://www.garlic.com/~lynn/submain.html#dasd
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Thu, 10 Feb 2011 09:55:53 -0500Peter Flass <Peter_Flass@Yahoo.com> writes:
the software has small profit margin (in the 60s, it use to be free) ... however, user being able to run an application program was the justification for having the hardware. Amdahl gave a talk in the early 70s at MIT about founding his clone processor company. He was asked how he convinced the venture/money people to fund the company. He said that customers had already invested several hundred billion in 360 application software development ... which even if IBM were to totally walk away from 360 (I claim could be a vieled reference to Future System), that was enough to keep him in business through the end of the century.
DOS to OS conversion was an incompatibility inhibitor ... but not as great as totally different system. Part of the issue was perception ... part of the issue was the degree of effort to move from DOS to OS (compared to radically different systems and hardware) ... aka possibly reduces the benefit of unified architecture ... but didn't totally defeat the benefit of unified architecture.
Low-end (actually any) customers could run an OS that eats 75% or more of his system ... if the aggregate cost of the system is less than the cost of not running the application (net benefit) ... and they didn't perceive a viable alternative. In early days it was completely different world ... cost of application development could dominate all other costs ... and opportunity costs ... having the automation versus doing it manually ... covered a lot of sins ... especially with high value operations like financial operations (financial institutions and/or financial operations of large corporations).
Compatible architecture provided perception that the several hundred
billion in (customer) software application development didn't have to
scrapped and start all over every generation. This was the testimony in
gov. litigation by RCA(?) that all the vendors had realized by the late
50s the requirement for single compatible line ... and only IBM actually
pulled it off (corollary is that with the perception of being the only
one that pulled it off, a lot of other things could be done wrong
... and still be able to dominate the market).
https://www.garlic.com/~lynn/2011b.html#57 If IBM Hadn't Bet the Company
POK favorite son operating system ... after FS failure and mad rush to
get product (hardware & software) back into the 370 line ... managed to
convince corporate to kill off vm370 (because POK needed all the vm370
developers in order to meet the MVS/XA delivery date). Endicott
eventually managed to save the vm370 product .. but had to reconstitute
a development group from scratch. misc. past posts mentioning FS
https://www.garlic.com/~lynn/submain.html#futuresys
some of the 70s confusion was that FS was planned to completely replace all 370 products (hardware and operating system) ... the aftermath of FS demise left the company in disarray and scrambling to recover (there have been comments if something the magnitude of FS had been attempted by any other vendor ... the resulting massive failure would have resulted in the vendor no longer being in business)
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Thu, 10 Feb 2011 14:56:41 -0500re:
company did have a large variety of other processors (besides 360/370)
... 1800, s/3, s/32, s/34, s/36, s/38, series/1, system/7, 8100,
etc. in addition there was large variety of embedded processors in
controller, devices, and the native engines for the low-end 370. old
email (joke) about MIT lisp machine project asking for 801/risc
processor and being offered 8100 instead:
https://www.garlic.com/~lynn/2003e.html#email790711
370 compatibility was already giving up a lot in performance, low&mid range emulators typically ran ratio of 10:1 native instructions to 370; 300kip 370/145 needed 3mip native engine, 80 kip 370/115 needed nearly 1mip native engine.
the 135/145 follow-on (370 138/148) had additional microcode storage
... plan was to 1) add additional (virtual machine) 370
privilege instructions to be executed directly to virtual
machine rules (rather than interrupt into kernel and be simulated
... aka "VMA" originally done for 370/158) and 2) "ECPS" which placed
part of the cp kernel directly in microcode. There was 6000 bytes of
extra microcode for "ECPS" and 370->native translated nearly
byte-for-byte. So the first effort was instrument the vm370 kernel and
identify the highest used kernel instruction pathlengths ... and
cut-off when 6000 bytes was reached (which accounted for just under of
80% of kernel overhead). The process involved invent a new instruction
and adding it to the vm370 kernel in front of the instruction sequence
it was replacing (with a pointer to the next following instruction
where it would resume). At startup, there was a test if the
appropriate microcode was available ... and if not, overlay all the
"ECPS" instructions with "no-ops". Old post with result of the vm370
kernel instrumentation that selected the kernel paths that would
become ECPS (part of work I did for Endicott spring of 1975):
https://www.garlic.com/~lynn/94.html#21 370 ECPS VM microcode assist
That 80% of the kernel pathlength (ECPS) then ran ten times faster (implemented in native engine) ... 370/148 at over 500kips ... ECPS then ran over 5mips (directly on the native engine).
Now, somewhat as result of complexity of Future System ... I've claimed that the work on 801/risc went to the opposite for extremely simple hardware. There was an advanced technology conference held in POK in the mid-70s ... where we presented 16-way 370 multiprocessor and the 801 group presented risc, CP.r and PL.8. End of 70s, a 801/risc effort was kicked off to replace the large variety of internal microprocessors with 801/risc (eliminating huge duplication in chip designs along with native engine software & application development). Low&mid range 370s would all converge to 801/risc as native microprocessor (4341 follow-on, the 4381 was suppose to be 801 native; s/38 follow-on, the as/400 was suppose to be 801 native, OPD displaywriter follow-on was to be 801 native, large number of microprocessors for controllers and other embedded devices would all be 801). For various reasons, these efforts all floundered and were aborted.
In spring 1982, I held internal corporate advance technology
conference (first since the earlier one held in POK, comments were that
in the wake of Future System demise ... most advance technology
efforts were all eventually scavanged to resume 370 activity and push
370 hardware and software out the door as quickly as possible).
https://www.garlic.com/~lynn/96.html#4a
I wanted to do a project that reworked vm/cms with lots of new
function ... implemented in higher level language ... to minimize it
being able to port to different machine architectures. In theory, this
would allow also being able to retarget a 370 kernel to some native
engine (possibly 801) while still providing both native virtual
machines and at the same time 370 virtual machines (best of ECPS
combined with being able to port to other processors) ... slight
analogy might be Apple with CMU MACH and power/pc. misc. old email
mentioning 801, risc, iliad, romp, rios, power, power/pc, etc
https://www.garlic.com/~lynn/lhwemail.html#801
note there had been big scavenging of advance technology towards
the end of FS (attempting to save the effort) ... reference in this
old email
https://www.garlic.com/~lynn/2011b.html#email800117
in this (linkedin) Future System post
https://www.garlic.com/~lynn/2011b.html#72
but then (after demise of FS) the mad rush to try and get stuff back into the 370 product pipelines ... they sucked nearly all remaining advanced technology resources into near term product production.
Other interests got involved (in SCP redo) and eventually the effort
was redirected to a stripped down kernel that could be common across
all the 370 operating systems. The prototype/pilot was to take the
stripped down tss/370 kernel ... which had been done for special
effort for AT&T ... they wanted to scaffold unix user interface on top
of the stripped down tss/370 kernel. The justification was with four
370 operating systems ... there was enormous duplication effort
required by all the device product houses ... having four times the
cost for device drivers and RAS support in the four different
operating systems (DOS, VS1, VM, and MVS). This became strategic
corporate activity (common low-level code for all four operating
systems) ... huge numbers of people assigned to it ... and eventually
had its own FS-demise moment ... collapsing under its own
weight. recent post on the subject
https://www.garlic.com/~lynn/2011.html#20
older reference about some of the TSS/370 analysis (for above)
https://www.garlic.com/~lynn/2001m.html#53
with all that going on ... by the mid-80s ... there were starting to
be single chip 370 implementations ... as well as various other/faster
801/risc chips ... so going in slightly different direction was large
clusters of processors (in rack mount configurations):
https://www.garlic.com/~lynn/2004m.html#17
also mentioned in this recent post
https://www.garlic.com/~lynn/2011b.html#48
and this collections of old email about NSFNET backbone ... I had to
find a fillin for presentation to head of NSF on HSDT ... because I
needed to do a week in YKT on the cluster processor stuff
http://lnkd.in/JRVnGk
other old NSFNET backbone related email
https://www.garlic.com/~lynn/lhwemail.html#nsfnet
and misc. old HSDT related email
https://www.garlic.com/~lynn/lhwemail.html#hsdt
the above cluster processor effort somewhat stalled ... but resumed
with medusa (cluster in a rack) ... some old email
https://www.garlic.com/~lynn/lhwemail.html#medusa
old referenced in this old post about jan92 meeting in Ellison's
conference room on cluster scale-up
https://www.garlic.com/~lynn/95.html#13
as part of ha/cmp product some old posts
https://www.garlic.com/~lynn/subtopic.html#hacmp
for other drift ... within a month of the meeting in Ellison's conference room, the cluster scale-up effort had been transferred, announced as supercomputer and we were told we couldn't work on anything with more than four processors. This contributed to decision to leave not long afterwards.
now a year or so later, two of the other people that were in that jan92 meeting ... had also left (oracle) and show up at a small client/server startup responsible for something called the "commerce server" ... the startup had also invented this technology they called "SSL". We were brought in to consult because they wanted to do payment transaction on the server; the result is now frequently called "electronic commerce".
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: NASA proves once again that, for it, the impossible is not even difficult. Newsgroups: comp.arch, comp.arch.embedded Date: Thu, 10 Feb 2011 15:10:31 -0500Robert Myers <rbmyersusa@gmail.com> writes:
there was a parody of the shuttle disaster with the booster rockets and the "o-rings" ... while there was a lot of attention payed to the operational characteristics of the "o-rings" ... the parody was the only reason that o-rings were required at all ... was because congress mandated that the booster rockets to be built near the rockies ... and then transported to the cape (requiring them to be in sections for transportation; resulting in the o-rings when they were assembled).
the parody was that somebody in the queen's court convinced her that columbus's ships had to be built in the mountains (where the trees were), then sawed into three pieces for transportation to the harbor ... and glued back together and then launched (as opposed to transporting the trees from the mountains to the harbor for construction of the ships). enormous resources then were focused on technology of gluing a ship back together after it had been sawed into three pieces (as opposed to deciding that ships could be built in the harbor and avoid having to saw them into pieces at all).
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: The first personal computer (PC) Newsgroups: alt.usage.english, alt.folklore.computers Date: Thu, 10 Feb 2011 19:28:32 -0500Michael Wojcik <mwojcik@newsguy.com> writes:
the virtual machine part provided for a kind of personal computers/computing.
I've made some number of references to it being the earlier generation of cloud computing.
misc. past posts referring to virtual machine based, commercial, online
timesharing
https://www.garlic.com/~lynn/submain.html#timeshare
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Thu, 10 Feb 2011 21:40:37 -0500"Joe Morris" <j.c.morris@verizon.net> writes:
and was picked up and shipped in the standard cp67 product.
there was lots of simplification in the morph from cp67 to vm370 ... and
all the fastpath code was dropped. Even tho i still hadn't done
port of cambridge system changes from cp67 to vm370 ... reference
here in this old email
https://www.garlic.com/~lynn/2006v.html#email731212
https://www.garlic.com/~lynn/2006w.html#email750102
https://www.garlic.com/~lynn/2006w.html#email750430
i did give the development group the "fastpath" changes that (i believe) were shipped in vm370 release 1, plc9.
justification for not restoring floating point registers in "fast redispatch" (interrupt into the kernel and resuming execution of the same virtual machine) was that cp kernel never used floating point registers (so they should have been unchanged during kernel execution).
multiprocessor support required getting smarter about "fast redispatch" (was the same virtual machine being redispatched on the same processor with contents of the real floating point registers still current for that virtual machine).
various fiddling in dispatch can result in all sorts of problems ...
here is recent (linkedin) post about fix "vm13025" ...
https://www.garlic.com/~lynn/2011b.html#61
including in above, old email with melinda:
https://www.garlic.com/~lynn/2011b.html#email860217
https://www.garlic.com/~lynn/2011b.html#email860217b
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Fri, 11 Feb 2011 00:50:19 -0500hancock4 writes:
sjr did a number of enhancements ... one was to RSCS to print random quotations on separator page (from the alternate paper drawer ... which was typically filled with colored paper (to make print job seperation easier).
another enhancement was all-points-addressable ... aka apa/sherpa.
old email mentioning sherpa
https://www.garlic.com/~lynn/2006p.html#email820304
in this post
https://www.garlic.com/~lynn/2006p.html#44 Materiel and graft
misc. other past posts mentioning apa/sherpa
https://www.garlic.com/~lynn/2005f.html#48 1403 printers
https://www.garlic.com/~lynn/2006p.html#49 Materiel and graft
https://www.garlic.com/~lynn/2006q.html#1 Materiel and graft
https://www.garlic.com/~lynn/2007g.html#27 The Complete April Fools' Day RFCs
https://www.garlic.com/~lynn/2007u.html#72 Parse/Template Function
https://www.garlic.com/~lynn/2008d.html#51 It has been a long time since Ihave seen a printer
https://www.garlic.com/~lynn/2008o.html#68 Blinkenlights
https://www.garlic.com/~lynn/2008o.html#69 Blinkenlights
https://www.garlic.com/~lynn/2010c.html#74 Apple iPad -- this merges with folklore
https://www.garlic.com/~lynn/2010e.html#43 Boyd's Briefings
https://www.garlic.com/~lynn/2010h.html#59 IBM 029 service manual
https://www.garlic.com/~lynn/2010k.html#49 GML
https://www.garlic.com/~lynn/2011.html#1 Is email dead? What do you think?
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Fri, 11 Feb 2011 09:43:57 -0500"Esra Sdrawkcab" <admin@127.0.0.1> writes:
There was (internal) email client called VMSG ... an extremely early 0.0 something pre-release version was used for core email in PROFS ... but they claimed to have done it themselves. Later when the author offered to provide them with the latest version with lots of upgrade ... PROFS attempted to get him fired. He managed to show that every PROFS email in the world carried his initials in a non-displayed field. After that he limited source distribution to only two other people (me and one other).
recent reference to the SNA/VTAM group had told members
of the executive committee that one of the reasons that
the internal network had to be converted to SNA/VTAM was
because PROFS was a VTAM application
https://www.garlic.com/~lynn/2011b.html#65 If IBM Hadn't Bet the Company
another reference to SNA/VTAM group working hard to
get internal network converted to SNA/VTAM
https://www.garlic.com/~lynn/2011.html#email870306
in this post
https://www.garlic.com/~lynn/2011.html#4 Is email dead? What do you think?
then reference to various internal factions (including SNA/VTAM) giving
me lots of problems working with NSF on the NSFNET backbone
http://lnkd.in/JRVnGk
including this old email reference (somebody had put together large
collection of emails with lots of misinformation trying to position
SNA/VTAM as solution for NSFNET backbone):
https://www.garlic.com/~lynn/2006w.html#email870109
in this posts
https://www.garlic.com/~lynn/2006w.html#21 SNA/VTAM for NSFNET
recent reference to author of VMSG also wrote PARASITE/STORY:
https://www.garlic.com/~lynn/2011b.html#67 If IBM Hadn't Bet the Company
misc. other past posts mentioning PROFS (&/or VMSG):
https://www.garlic.com/~lynn/99.html#35 why is there an "@" key?
https://www.garlic.com/~lynn/2000c.html#46 Does the word "mainframe" still have a meaning?
https://www.garlic.com/~lynn/2000e.html#20 Is Al Gore The Father of the Internet?^
https://www.garlic.com/~lynn/2001j.html#35 Military Interest in Supercomputer AI
https://www.garlic.com/~lynn/2001k.html#35 Newbie TOPS-10 7.03 question
https://www.garlic.com/~lynn/2001k.html#39 Newbie TOPS-10 7.03 question
https://www.garlic.com/~lynn/2001k.html#40 Newbie TOPS-10 7.03 question
https://www.garlic.com/~lynn/2001k.html#56 E-mail 30 years old this autumn
https://www.garlic.com/~lynn/2002f.html#14 Mail system scalability (Was: Re: Itanium troubles)
https://www.garlic.com/~lynn/2002h.html#58 history of CMS
https://www.garlic.com/~lynn/2002h.html#59 history of CMS
https://www.garlic.com/~lynn/2002h.html#64 history of CMS
https://www.garlic.com/~lynn/2002i.html#50 CDC6600 - just how powerful a machine was it?
https://www.garlic.com/~lynn/2002j.html#4 HONE, ****, misc
https://www.garlic.com/~lynn/2002p.html#34 VSE (Was: Re: Refusal to change was Re: LE and COBOL)
https://www.garlic.com/~lynn/2003b.html#45 hyperblock drift, was filesystem structure (long warning)
https://www.garlic.com/~lynn/2003e.html#69 Gartner Office Information Systems 6/2/89
https://www.garlic.com/~lynn/2003j.html#56 Goodbye PROFS
https://www.garlic.com/~lynn/2003m.html#26 Microsoft Internet Patch
https://www.garlic.com/~lynn/2004j.html#33 A quote from Crypto-Gram
https://www.garlic.com/~lynn/2004p.html#13 Mainframe Virus ????
https://www.garlic.com/~lynn/2005t.html#43 FULIST
https://www.garlic.com/~lynn/2005t.html#44 FULIST
https://www.garlic.com/~lynn/2006d.html#10 IBM 610 workstation computer
https://www.garlic.com/~lynn/2006n.html#23 sorting was: The System/360 Model 20 Wasn't As Bad As All That
https://www.garlic.com/~lynn/2006q.html#4 Another BIG Mainframe Bites the Dust
https://www.garlic.com/~lynn/2006t.html#42 The Future of CPUs: What's After Multi-Core?
https://www.garlic.com/~lynn/2006w.html#42 vmshare
https://www.garlic.com/~lynn/2006x.html#7 vmshare
https://www.garlic.com/~lynn/2007b.html#14 Just another example of mainframe costs
https://www.garlic.com/~lynn/2007b.html#31 IBMLink 2000 Finding ESO levels
https://www.garlic.com/~lynn/2007b.html#32 IBMLink 2000 Finding ESO levels
https://www.garlic.com/~lynn/2007d.html#17 Jim Gray Is Missing
https://www.garlic.com/~lynn/2007e.html#4 The Genealogy of the IBM PC
https://www.garlic.com/~lynn/2007f.html#13 Why is switch to DSL so traumatic?
https://www.garlic.com/~lynn/2007j.html#50 Using rexx to send an email
https://www.garlic.com/~lynn/2007p.html#29 Newsweek article--baby boomers and computers
https://www.garlic.com/~lynn/2007v.html#54 An old fashioned Christmas
https://www.garlic.com/~lynn/2007v.html#55 An old fashioned Christmas
https://www.garlic.com/~lynn/2007v.html#63 An old fashioned Christmas
https://www.garlic.com/~lynn/2008.html#69 Rotary phones
https://www.garlic.com/~lynn/2008.html#75 Rotary phones
https://www.garlic.com/~lynn/2008h.html#46 Whitehouse Emails Were Lost Due to "Upgrade"
https://www.garlic.com/~lynn/2008k.html#59 Happy 20th Birthday, AS/400
https://www.garlic.com/~lynn/2009.html#8 Is SUN going to become x86'ed ??
https://www.garlic.com/~lynn/2009.html#23 NPR Asks: Will Cloud Computing Work in the White House?
https://www.garlic.com/~lynn/2009k.html#0 Timeline: The evolution of online communities
https://www.garlic.com/~lynn/2009k.html#16 Mainframe hacking
https://www.garlic.com/~lynn/2009l.html#41 another item related to ASCII vs. EBCDIC
https://www.garlic.com/~lynn/2009m.html#34 IBM Poughkeepsie?
https://www.garlic.com/~lynn/2009o.html#33 U.S. house decommissions its last mainframe, saves $730,000
https://www.garlic.com/~lynn/2009o.html#38 U.S. house decommissions its last mainframe, saves $730,000
https://www.garlic.com/~lynn/2009q.html#43 The 50th Anniversary of the Legendary IBM 1401
https://www.garlic.com/~lynn/2009q.html#49 The 50th Anniversary of the Legendary IBM 1401
https://www.garlic.com/~lynn/2009q.html#51 The 50th Anniversary of the Legendary IBM 1401
https://www.garlic.com/~lynn/2009q.html#64 spool file tag data
https://www.garlic.com/~lynn/2009q.html#66 spool file data
https://www.garlic.com/~lynn/2010.html#1 DEC-10 SOS Editor Intra-Line Editing
https://www.garlic.com/~lynn/2010b.html#8 Happy DEC-10 Day
https://www.garlic.com/~lynn/2010b.html#44 sysout using machine control instead of ANSI control
https://www.garlic.com/~lynn/2010b.html#87 "The Naked Mainframe" (Forbes Security Article)
https://www.garlic.com/~lynn/2010b.html#96 "The Naked Mainframe" (Forbes Security Article)
https://www.garlic.com/~lynn/2010b.html#97 "The Naked Mainframe" (Forbes Security Article)
https://www.garlic.com/~lynn/2010c.html#88 search engine history, was Happy DEC-10 Day
https://www.garlic.com/~lynn/2010d.html#61 LPARs: More or Less?
https://www.garlic.com/~lynn/2010o.html#4 When will MVS be able to use cheap dasd
https://www.garlic.com/~lynn/2010q.html#45 Is email dead? What do you think?
https://www.garlic.com/~lynn/2011.html#4 Is email dead? What do you think?
https://www.garlic.com/~lynn/2011b.html#10 Rare Apple I computer sells for $216,000 in London
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Fri, 11 Feb 2011 10:32:04 -0500Roland Hutchinson <my.spamtrap@verizon.net> writes:
from above:
The IBM 3800 laser-electrophotographic printer of 1975 had a speed of
20,000 lines a minute in preparing bank statements, premium notices and
other high-volume documents. Laser beam paths were altered millions of
times a second and were reflected from an 18-sided mirror that spun at
12,000 revolutions per minute.
... snip ...
laser printer wiki
https://en.wikipedia.org/wiki/Laser_printer
re:
https://www.garlic.com/~lynn/2011b.html#82 If IBM Hadn't Bet the Company
google usenet from 1985 mentioning XEROX laser printer emulating 6670.
http://groups.google.com/group/fa.laser-lovers/browse_thread/thread/bb59533ee5d75b2d
OPD history ... mentioning a number of things in the 60s & 70s,
including "Copier II"
http://www-03.ibm.com/ibm/history/exhibits/modelb/modelb_office2.html
it then mentions in 1976 ... Series III Copier/Duplicator (which I presume is same as "Copier III" ... This was essentially what the 6670 derived from with computer interface).
reference to 6670:
http://www-03.ibm.com/ibm/history/reference/faq_0000000011.html
above also mentions powerparallel SP2 with up to 128 nodes announced April 1994.
from these posts:
https://www.garlic.com/~lynn/2001n.html#70
https://www.garlic.com/~lynn/2001n.html#83
earlier version was SP1 ... referenced in these press items:
https://www.garlic.com/~lynn/2001n.html#6000clusters2 11May92
and
https://www.garlic.com/~lynn/2001n.html#6000clusters1 2/17/92
above was just barely month later than this Jan92 meeting in
Ellison's conference room
https://www.garlic.com/~lynn/95.html#13
other old email mentioning cluster scale-up
https://www.garlic.com/~lynn/lhwemail.html#medusa
as part of ha/cmp product effort
https://www.garlic.com/~lynn/subtopic.html#hacmp
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: The first personal computer (PC) Newsgroups: alt.folklore.computers Date: Fri, 11 Feb 2011 10:47:30 -0500Anne & Lynn Wheeler <lynn@garlic.com> writes:
this mentions "leland stanford junior varsity farm":
http://www.thehulltruth.com/dockside-chat/17756-college-b-ball-hawks-zags-et-al.html
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Fri, 11 Feb 2011 11:21:04 -0500hancock4 writes:
one of the big business pushes is to take the call-center support ... many with large rooms of terminals run by CICS ... and do "webified" front end (for the CICS screens, some of this could just be the old-time "screen scrapping" that had been done on PCs with HLLAPI terminal emulation application) ... with various security, authentication, and other constraints ... allowing end-users to do a lot of their own operations ... w/o having to involve call-center.
a decade ago, i visited a datacenter that had large banner on the wall with something about over 120 CICS "regions". this is before CICS had multiprocessor support ... so method to increase concurrent operations was to operate multiple "regions", i.e. independent copies/invokations of CICS. CICS had its own highly optimized multithreaded support ... but (each CICS) appeared as a single (processor), serialized execution to the operating system.
the datacenter claimed to provide dataprocessing for nearly all the cable tv companies in the US ... all the local cable tv office terminal screens were CICS into this datacenter, it did all the billing and accounting ... and backend for cable tv call-centers ... it also was the "head-end" for all the cable tv settop boxes ... sending out signals that eventually changed/updated individual settop boxes.
reference to CICS finally getting multiprocessor support in 2004
https://web.archive.org/web/20041023110006/http://www.yelavich.com/history/ev200402.htm
where you could get simultaneous execution threads running concurrently in multiple processors for the same CICS image (although it required a new application conformace for "threadsafe").
misc. past posts mentioning CICS (&/or BDAM):
https://www.garlic.com/~lynn/submain.html#cics
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: The first personal computer (PC) Newsgroups: alt.usage.english, alt.folklore.computers Date: Fri, 11 Feb 2011 11:36:58 -0500Walter Bushell <proto@panix.com> writes:
they obvious didn't have my fairshare scheduler ... that i had
originally done for cp67 as undergraduate (it actually established
generalized resource consumption policies ... with the default being
"fairshare")
https://www.garlic.com/~lynn/subtopic.html#fairshare
things would still degrade under heavy load ... but much more gracefully ... and trivial interactive activity tended to be fairly well insulated from heavy users.
This was dropped in part of the simplification morph from cp67 to vm370
... but I got later re-introduce it as the "resource manager". However,
this was after the future system demise and the mad rush to get products
back into the 370 hardware&software product pipelines (contributing to
decided to release stuff that I had been doing all during the future
system period) ... misc. past posts mentioning "future system"
https://www.garlic.com/~lynn/submain.html#futuresys
the distraction of future system (and future system doing its best to kill off all 370 activity ... as part of supporting their strategic position) ... is claimed as contributing to clone processors getting market foothold.
the 23jun69 unbundling announcement (in response to various litigation)
had started charging for software (and other changes) ... but the
company managed to make the case (with the gov) that kernel software
should still be free.
https://www.garlic.com/~lynn/submain.html#unbundle
however, later with the clone processors in the market, the decision was
made to transition to charging for kernel software ... and my resource
manager was selected as guinea pig ... which met that I to spend a lot
of time with business & legal people on policies for kernel software
charging. the transition period also had other implications
... discussed in this recent long-winded post:
https://www.garlic.com/~lynn/2011b.html#61 VM13025 ... zombie/hung users
During the transition, there was combination of free & charged-for kernel software ... with sometimes peculiar interaction ... after several years ... the transition was complete and all software was being charged for. About that time, the OCO-wars (object code only) began, in addition to charging for all software ... also no longer provided full source.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: NASA proves once again that, for it, the impossible is not even difficult. Newsgroups: comp.arch, comp.arch.embedded Date: Fri, 11 Feb 2011 12:47:04 -0500Paul Colin Gloster <Colin_Paul_Gloster@ACM.org> writes:
superfreakonomics has a bit on how cities of the world had much more severe polution problem before the advent of automobiles and internal combustion engine ... and that NYC had higher rate of traffic deaths per thousand from the horse era than they now have from automobiles.
it also explained why the brownstones in NYC where so high above the ground (with front steap steps) ... because of the horse manure piled so high on the streets.
--
virtualization experience starting Jan1968, online at home since Mar1970
From: Anne & Lynn Wheeler <lynn@garlic.com> Subject: Re: If IBM Hadn't Bet the Company Newsgroups: alt.folklore.computers Date: Fri, 11 Feb 2011 15:02:40 -0500re:
from long ago and far away
Date: 18 May 1984, 09:59:01 PDT
To: distribution
Subject: Floating Point Registers corrupted by HPO 2.5 and above
You may have already been informed of this problem, but I wanted to
make sure everyone in the area had the details.
APAR VM20536 describes a problem with all releases of CP from HPO 2.5
on up. Under certain circumstances it is possible for a virtual
machine to be dispatched with the wrong set of floating point registers -
a major annoyance, I'm sure you'll agree. For HPO 3.4 the fix for this
APAR has a pre-requisite of VM20519.
Information from RETAIN on VM20536 follows:
PIN 5749-DM-K00-952 VM20536-IN-INCORROUT ACTIVE F999- -S/FIX- OPTIONS ARE: CMS PAGE FAULTS ON LOAD OF FLOATING POINT REGISTERS PRINT... snip ... top of post, old email index
END OF ABSTRACT FESN6401302- REPORTED RELEASE RD95 ERROR DESCRIPTION: CMS VIRTSYS TAKES A PAGE FAULT AFTER A LOAD FPRS INSTRUCTION. AFTER THE CMS MACHINE IS DISPATCHED TO RUN, THE FPRS ARE RESTORED INCORRECTLY.
PROBLEM SUMMARY: USERS AFFECTED: ALL HPO2.5, HPO3, HPO3.2 AND HPO3.4 USERS FULL PAGE= ON FAST PATH DISPATCH FAILS TO RELOAD FLOATING POINT REGISTERS. PIN PROBLEM CONCLUSION: PG 1, 2 OF 6 DMKDSP WILL BE CHANGED TO RELOAD FLOATING POINT REGS CORRECTLY. TEMPORARY FIX: FIX IS AVAILABLE VIA A PTF REQUEST APAR= VM20536 OWNED BY:
the following had long list of MVC & CLC instructions in DMKDSP ... included one that was inserted as part of 20536 fix for floating point fast redispatch problem (I've snipped/pruned the list):
Date: 7 May 1986, 22:46:16 EDT
To: distribution
Subject: DMKDSP Performance
The following instructions are from the current version of DMKDSP (HPO
4.2) These are the 2 slowest instructions in the 370 BAL. Its stupid to
have them in the most heavily used module in the system.
< ... snip ... >
CLC VMLSTPRC,LPUADDR+1 WAS VM LAST RUN ON THIS PROC @VA20536 29810920
CLC VMLSTPRC,LPUADDR+1 WAS VM LAST RUN ON THIS PROC @V407508 29940000
CLC VMLSTPRC,LPUADDR+1 WAS VM LAST RUN ON THIS PROC %VC5QAN0 31563000
... snip ... top of post, old email index
DMKPRV Assembler (1/31/74) VM/370 Release 2 with no updates:
SETREAL EQU * SET REAL STORAGE KEY 0059800
TM VMPSTAT,VMREAL V=R USER ? @VA01071 0059810
BO *+8 YES - LEAVE ALL OF KEY "AS IS" @VA01071 0059820
N R9,=A(X'F8') MASK FOR REAL KEY (WITH @VA01071 0059910
• FETCH-PROT. BIT) 0059920
TM SWPFLAG,SWPSHR SHARED SYSTEM PAGE ? @VA01071 0060000
BCR 1,R10 YES - JUST LEAVE 0060100
LRA R2,0(0,R6) SEE IF PAGE IN CORE 0060200
BCR 7,R10 PAGE NOT IN CORE 0060300
SSK R9,R2 SET REAL KEY 0060400
TM VMOSTAT,VMSHR RUNNING SHARED-SEGMENT SYSTEM ? 0060500
BCR 8,R10 NO -- JUST RE-DISPATCH 0060600
LA R15,240 CHECK FOR KEY OF 0 @VA01071 0060710
NR R15,R9 DO NOT SET REAL ZERO KEY @VA01071 0060720
BCR 7,R10 <BNZ> IF NON-ZERO, IT'S OK. @VA01071 0060730
-----------------------------------------------------------------------
DMKPRV Assembler VM/370 Release 4 with HPO 4.2 updates:
*---------------------------------------------------------------------- 1125621
• HANDLE SIMULATION OF THE SSK INSTRUCTION. 1125690
*---------------------------------------------------------------------- 1125759
SPACE 1 1125828
DOSSK EQU * @V6KT2LD 1125897
IC R9,VMGPRS+3(R5) GET NEW KEY 1126000
NR R9,R8 MASK FOR USER 1127000
STC R9,SWPKEY1(R1) SET IN SWPTABLE ENTRY @VA01071 1128000
SETREAL EQU * SET REAL STORAGE KEY 1129000
C R11,AVMREAL IS THIS THE V=R USER? @VA12156 1130200
BNE NVER NO,THEN BUSINESS AS USUAL @VA12156 1130400
TM CPSTAT2,CPSPMODE ARE WE IN SPMODE? @VA12156 1130600
BZ NVER1 NO,THEN BUSINESS AS USUAL @VA12156 1130800
SLR R2,R2 CLEAR R2 FOR SWPTABLE UPDATE @VA12156 1131000
STC R2,SWPKEY1(R1) PUT SWPTABLE KEY TO ZERO FOR @VA12156 1131200
• V=R USER SO CC IS CORRECT FOR 1131400
• VMA IN SPMODE. 1131600
B NVER1 @VA12156 1131800
NVER DS 0H @VA12156 1132000
N R9,=A(X'F8') V=V MASK FOR KEY,FETCH-PROT-BIT @VA01071 1132399
NVER1 DS 0H @VA12156 1132600
LRA R2,0(0,R6) SEE IF PAGE IN CORE 1134000
BCR 7,R10 PAGE NOT IN CORE 1135000
SSK R9,R2 SET REAL KEY 1136000
BR R10 FAST DISPATCH 1137000
... snip ... top of post, old email index
old items mentioning moving off of CP67 to VM370:
https://www.garlic.com/~lynn/2006v.html#email731212
https://www.garlic.com/~lynn/2006w.html#email750102
https://www.garlic.com/~lynn/2006w.html#email750430
also as I've previously mentioned it wasn't long afterwards that Almaden
had operations problem with random tapes being mounted for scratch
... and I lost enormous amount of stuff ... even things that had been
replicated on 2-3 different tapes. past posts mentioning Almaden tape
library operational problems:
https://www.garlic.com/~lynn/2003j.html#14 A Dark Day
https://www.garlic.com/~lynn/2006w.html#42 vmshare
https://www.garlic.com/~lynn/2007l.html#51 Scholars needed to build a computer history bibliography
https://www.garlic.com/~lynn/2008q.html#52 TOPS-10
https://www.garlic.com/~lynn/2009.html#8 Is SUN going to become x86'ed ??
https://www.garlic.com/~lynn/2009.html#13 Is SUN going to become x86'ed ??
https://www.garlic.com/~lynn/2009d.html#4 Why do IBMers think disks are 'Direct Access'?
https://www.garlic.com/~lynn/2009f.html#59 Backup and Restore Manager for z/VM
https://www.garlic.com/~lynn/2009n.html#66 Evolution of Floating Point
https://www.garlic.com/~lynn/2009s.html#17 old email
https://www.garlic.com/~lynn/2010.html#4 360 programs on a z/10
https://www.garlic.com/~lynn/2010b.html#51 Source code for s/360
https://www.garlic.com/~lynn/2010b.html#96 "The Naked Mainframe" (Forbes Security Article)
https://www.garlic.com/~lynn/2010d.html#65 Adventure - Or Colossal Cave Adventure
https://www.garlic.com/~lynn/2010l.html#0 Old EMAIL Index
https://www.garlic.com/~lynn/2010q.html#45 Is email dead? What do you think?
https://www.garlic.com/~lynn/2011b.html#39 1971PerformanceStudies - Typical OS/MFT 40/50/65s analysed
--
virtualization experience starting Jan1968, online at home since Mar1970