2003 Newsgroup postings (02/16 - 03/10) Lynn Wheeler

2003 Newsgroup Postings (02/16 - 03/10)

big buys was: Tubes in IBM 1620?
iso 8583
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
New England Winter Drivers
iso 8583
Surprising discovery
Low-end processors (again)
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
Low-end processors (again)
osi archeological reference from ietf-pkix list
difference between itanium and alpha
COMTEN- IBM networking boxes
OT: Attaining Perfection
CA-RAMIS
Card Columns
CA-RAMIS
Efficent Digital Signature Schemes
PC history, was PDP10 and RISC
COMTEN- IBM networking boxes
PDP10 and RISC
Which Editor
CPU Impact of degraded I/O
CPU Impact of degraded I/O
Which Editor
Antiquity of Byte-Word addressing?
[urgent] which OSI layer is SSL located?
Why only 24 bits on S/360?
SSL questions
SSL questions
Why only 24 bits on S/360?
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
Why only 24 bits on S/360?
YKYBHTLW
Why only 24 bits on S/360?
Why only 24 bits on S/360?
Why only 24 bits on S/360?
The PDP-1 - games machine?
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
Authentification vs Encryption in a system to system interface
Why only 24 bits on S/360?
Authentification vs Encryption in a system to system interface
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
Why only 24 bits on S/360?
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
unix
IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
The PDP-1 - games machine?
unix
unix
Antiquity of Byte-Word addressing?
PDP10 and RISC
Reviving Multics
Filesystems
Another light on the map going out
Another light on the map going out
Another light on the map going out
POWER hashes vs tree
unix
Bitnet again was: unix
Another light on the map going out
ARPAnet again: Bitnet again was: unix
ARPAnet again: Bitnet again was: unix
IBM was: VAX again: unix
unix
unix
unix
unix
unix
DOS trivia question
SSL/TLS DHE suites and short exponents
cp/67 35th anniversary
unix
unix
unix
reviving Multics
'Boyd': A military Strategist's Emphasis on Speed

big buys was: Tubes in IBM 1620?

Refed: **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: big buys was: Tubes in IBM 1620?
Newsgroups: alt.folklore.computers
Date: Sun, 16 Feb 2003 18:05:49 GMT

eugene@cse.ucsc.edu (Eugene Miya) writes:
That wasn't unique to 4341s. That happened with DEC VAX-11/780s and before that PDP-11s, PDP-8s, and IBM Series/1 machines. The only place you didn't see that was in high end computing. These mass purchases created havoc for smaller lot purchasers. That's why one site I worked for with an order for 300 minis got awarded to some obscure firm because DEC was too busy working on a 500 CPU order.

similar thread re: vax with world-wide ship numbers (thru '87):
https://www.garlic.com/~lynn/2002f.html#0 Computers in Science Fiction
https://www.garlic.com/~lynn/2002i.html#20 CDC6600 - just how powerful a machine was it

one of the things we looked at when we were doing ha/cmp
https://www.garlic.com/~lynn/2002f.html#1 Blade architctures

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

iso 8583

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: iso 8583
Newsgroups: alt.technology.smartcards
Date: Sun, 16 Feb 2003 21:37:21 GMT

"Rohit" writes:
Why do u think there is so much work going on in coverting iso 8583 format to xml ?

I mean do u think it is going to solve any problem

it may be the reverse .... some number of people are looking at defining all sorts of xml infrastructures ... and it would appear that some number of these might include the concept of transfer of money or at least the ability to be able to generate a payment transactions. one of the straight-forward mechanisms would be to find out what elemants are required in 8583 payment transactions and make sure there is some mapping from the xml environment to the 8583 environment (aka making sure that an xml originated something or other included enuf of the necessary data elements to be able to generate something that passed for a valid 8583 transaction at some gateway).

an example might be the ECML (electronic commerce modeling language) standardization work going on in ietf/internet. they aren't trying to define an 8583 transaction in xml. I believe they do want to have something that looks and tastes like xml .... and include the ability to initiate funds transfer as part of the operation (for some sort of value eachange ... aka want to buy something and need to pay for it). to that extent, there could easily be some reality check to see if the ECML definition included sufficient field definitions/coverage from which a real 8583 payment transaction might be generated.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

Refed: **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
 monopoly
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Tue, 18 Feb 2003 01:40:59 GMT

bill@gw5.cs.uofs.edu (Bill Gunshannon) writes:
History of X

In 1984 the Massachusetts Institute of Technology (MIT) formed Project Athena. The goal was to take the existing assortment of incompatible workstations from different vendors and develop a network of graphical workstations that could be used as teaching aids. The solution was a network that could run local applications while being able to call on remote resources. They thus created the first operating environment that was truly hardware and vendor independent - the X Window System.

athena news letters from 84 & 85:
http://www.mit.edu/afs/athena/system/usrdoc/athena/newsletter/

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

Refed: **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
 monopoly
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Tue, 18 Feb 2003 02:04:36 GMT

koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:
I wonder who writees this stuff, the latest batch of kids? I recall trying to port around hardware and vendor independent graphics stuff from the 70s.

And X11 alone isn't an operating environment, its just a begining of a GUI.

you still had terminal rooms ... except in project athena case they eventually became workstation rooms (workstations were still too expensive for every student to have one) .... sort of half dec workstations and half (ibm) pc/rt. they were on a lan connected to various university servers. the concept was that you could walk up to any machine, sit down ... and with a few magic keystrokes you had your complete personalized computing environment.

see previous post with:
http://www.mit.edu/afs/athena/system/usrdoc/athena/newsletter/

as in the 10/1/84 newsletter:
http://www.mit.edu/afs/athena/system/usrdoc/athena/newsletter/84-10-01

the terminal rooms started out with pc/xt

and as in
http://www.mit.edu/afs/athena/system/usrdoc/athena/newsletter/85-09

unix service start out with 45 vax 11/750 running unix timesharing. also in the above. there is an introduction to "The X Window System" by Win Treeese (getting started on a VS100). There is reference to more about x in:
http://www.mit.edu/afs/athena/system/usrdoc/athena/newsletter/85-05

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

New England Winter Drivers

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: New England Winter Drivers
Newsgroups: alt.folklore.computers
Date: Tue, 18 Feb 2003 01:44:26 GMT

hawk@slytherin.ds.psu.edu (Dr. Richard E. Hawkins) writes:
U.S. Marines, not Royal Marines :)

Each MRE is about 2 pounds, and these guys are sent out with 100+ lbs on their backs. They stripped down the meals and tossed everything but the high-carb stuff before they left. They also replaced their GI GPS systems with one from a sporting goods store that used the same batteries as something else in their packs, and made a few more modifications.

Much will be standard issue soon; the CO listened to the email from the sergeant, and passed it up . . .

somebody had a bright idea and had the knife upgraded with a compass in the end of the knife handle ... right where you use it as a hammer. the first thing to go was the compass.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

iso 8583

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: iso 8583
Newsgroups: alt.technology.smartcards
Date: Tue, 18 Feb 2003 05:40:27 GMT

"Rohit" writes:
So u think that xml could be a good substitute to iso 8583.

Morever i feel that it could end the proprietary and fixed format issues.

eliza, right?

normally eliza does things like ... why do you feel that way ... or why would you believe that.

the parser in this version of eliza doesn't seem to be able to understand the contradiction in calling an international standard, proprietary. the other part ... probably isn't really a parser problem ... not having been feed & understood iso8583 ... the parser wouldn't realize the contridction re the part about fixed format.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Surprising discovery

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Surprising discovery
Newsgroups: comp.arch
Date: Tue, 18 Feb 2003 15:20:25 GMT

Robert Myers writes:
Some dates:

1972 NASA F8 experimental fly-by-wire
1981 Boeing 767 computer-in-the-cockpit, partially fly-by-wire
1984 Airbus A320 fully fly-by-wire
1995 Boeing 777 fully fly-by-wire

drift, boyd's f16 ... 1976-1979
http://www.af.mil/news/factsheets/F_16_Fighting_Falcon.html
http://www.fas.org/man/dod-101/sys/ac/f-16.htm
http://www.f16falcon.com/facts/f16_1.html
http://www.csd.uwo.ca/~pettypi/elevon/baugher_us/f016.html
http://www.jetplanes.co.uk/f16.html

misc. boyd refs:
https://www.garlic.com/~lynn/subboyd.html#boyd

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Low-end processors (again)

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Low-end processors (again)
Newsgroups: bit.listserv.ibm-main,alt.folklore.computers
Date: Tue, 18 Feb 2003 16:22:45 GMT

IBM-MAIN@ISHAM-RESEARCH.COM (Phil Payne) writes:
Since systems aren't licensed by their I/O capacity and most VSE workloads have high I/O content, it seems that this is an area where vendors might compete on performance and thus where any benchmarks should be.

one might assert that just about all commercial application have been i/o bound since possibly the mid-70s ... and any apparent processor intensive activity might be a subject for some investigation.

discussions about relative dasd system perforaance declined by five to ten times over a period of 15 years (aka processor and memory increased by factor of fifty, dasd only increased by factor of five resulting in dasd relative system performance possibly declining by over an order of magnitude):
https://www.garlic.com/~lynn/93.html#31 Big I/O or Kicking the Mainframe out the Door
https://www.garlic.com/~lynn/95.html#10 Virtual Memory (A return to the past?)
https://www.garlic.com/~lynn/98.html#46 The god old days(???)
https://www.garlic.com/~lynn/99.html#4 IBM S/360
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/99.html#190 Merced Processor Support at it again
https://www.garlic.com/~lynn/2001f.html#62 any 70's era supercomputers that ran as slow as today's supercomputers?
https://www.garlic.com/~lynn/2001l.html#40 MVS History (all parts)
https://www.garlic.com/~lynn/2001l.html#61 MVS History (all parts)
https://www.garlic.com/~lynn/2001m.html#23 Smallest Storage Capacity Hard Disk?
https://www.garlic.com/~lynn/2002.html#5 index searching
https://www.garlic.com/~lynn/2002b.html#11 Microcode? (& index searching)
https://www.garlic.com/~lynn/2002b.html#20 index searching
https://www.garlic.com/~lynn/2002e.html#8 What are some impressive page rates?
https://www.garlic.com/~lynn/2002e.html#9 What are some impressive page rates?
https://www.garlic.com/~lynn/2002i.html#16 AS/400 and MVS - clarification please
https://www.garlic.com/~lynn/2002n.html#58 IBM S/370-168, 195, and 3033
https://www.garlic.com/~lynn/2003.html#21 vax6k.openecs.org rebirth

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
monopoly
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Tue, 18 Feb 2003 16:42:15 GMT

pechter@shell.monmouth.com (Bill/Carolyn Pechter) writes:
The OSF Logical Volume Manager still exists in IBM AIX...

most of OSF was technology provided by various members. distributed computer meetings were looking at merging stuff from (at least) MIT, CMU, AIX, Apollo, and UCLA's Locus.

LVM was originally developed as part of the original aixv3 for 6000 (along with JFS) .... and predated OSF.

aixv3 was sort of port of aixv2 from pc/rt with lots of changes (which was at&t derivative)

one of the prominant people that went to OSF had been manager of APL in STL ... and then went to PASC as manager of BSD-for-370 project ... which got retarged to BSD-for-PC/RT (and called AOS) ... and is now at Lotus.

another person i was sort of surprised to see at OSF (by way of DEC) was one of the old CSC/VM alumni who was one of the people that migrated to DEC when they closed the burlington mall location and transferred people to POK.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
monopoly
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Tue, 18 Feb 2003 20:09:05 GMT

JF Mezei writes:
I think that IBM truly changed mentality under Gerstner. IBM was nearly bankrupt and its "proprietary" mentality had to change. And by that, I didn't mean MVS or the 360 architecture, I meant the closed minded shop where disk drives didn't exist, tey were all DASD, ethernet didn't exist, it was all token ring or SNA etc. When IBM opened up and accepted established standards such as TCPIP , it gave it a new breath of life.

Microsoft was almost like that too. Remember that Gates was slow to see and adopt the emergence of the internet.

during the '80s ibm created IBUs (independent business units) ... acorn was an IBU
https://www.garlic.com/~lynn/2002g.html#79 Coulda, Woulda, Shoudda moments?

awd (brought you pc/rt, rs/6000, rios, power, aix, etc) was an IBU. there were some number of other IBUs.

also during the '80s, ACIS (academic) was formed with initial kitty of something like $300m to give away to educational institutions (and support things like NSFNET/internet).

part of the IBU problem was that it still drew on people with traditional large organization background. IBUs were supposedly given freedom to totally operate as an independent corporation .... but rarely did the traditional people stray very far from the corporate line. there was also some joke about an IBU claiming that they were free of some corporate hdqtrs bureaucacy .... but it turned out that each and every corporate hdqtrs bureaucrat would say that IBUs are free from all the other bureaucratic processes ... but not mine.

I believe that the company hit a peak in the range of 500k employees sometime in the '80s. it then went into the red in 1992 .... and by the time gerstner came on board employees were down in the range of 200k. as a result it was already a significantly different company ... although it still retained quite a few of the same top executives.

prior to gerstner coming on board ... there was lots of work being done breaking the company into independent corporations (ala the baby bells). there were also jokes about would the last person to leave, please turn out the lights. gerstner pretty much put a stop to the break up (although the spin-off of san jose to hitachi might only indicate that they were just suspended). The whole adstar name brand thing was getting ready for the split and the new disk company's name.

total drift was that workstation datasave went thru ADSM (adstar storage manager) and finally TSM (tivoli storage manager). Tivoli, itself started off as a couple austin awd employees leaving and doing a startup.

random adsm drift:
https://www.garlic.com/~lynn/2001n.html#66 Holy Satanism! Re: Hyper-Threading Technology - Intel information.
https://www.garlic.com/~lynn/2002e.html#3 IBM's "old" boss speaks (was "new")
https://www.garlic.com/~lynn/2002e.html#10 Deleting files and emails at Arthur Andersen and Enron
https://www.garlic.com/~lynn/2002h.html#29 Computers in Science Fiction
https://www.garlic.com/~lynn/2002k.html#8 Avoiding JCL Space Abends
https://www.garlic.com/~lynn/2002q.html#25 Beyond 8+3

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Low-end processors (again)

Refed: **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Low-end processors (again)
Newsgroups: bit.listserv.ibm-main,alt.folklore.computers
Date: Wed, 19 Feb 2003 16:09:24 GMT

BPerryman@TNSI.COM (Perryman, Brian) writes:
Since another 'mainframe' platform we have here is the high-availability Stratus Continuum running VOS, we standardised on their FTServer where we also later needed high-availability some NT-based POS/ATM transaction acquisition software (in fact this platform is now that particular software vendor's recommended hardware also).

It would make management here (and me) very happy if we could run our IBM 390 workload (OS/390 and/or z/OS) via Flex-ES on a Stratus FTServer..

http://www.stratus.com/products/ftserver/

Ok, it's not the cheapest Intel server around, but they're certainly HA. If anyone could comment on the suitability of these from a hardware spec or other point of view, I'd much appreciate it.

I definitely don't want to run MVS on a 'toy' - I'm looking to run a 24/7 production workload here, not play with something for academic interest.

when my wife and I were running ha/cmp .... we spent a lot of time working with the 88 Product Administrator (a re-logoed stratus). at that time stratus was purely fault tolerant and needed downtime for things like software maintenance. we had discussions with a number of places where ha was actually preferred to fault tolerant .... since ha also could mask software maintenance downtime. misc refs:
https://www.garlic.com/~lynn/95.html#13 SSA
https://www.garlic.com/~lynn/subtopic.html#hacmp

I also had something like a two hour running argument at a acm sigops meeting about whether it was possible to do real ha on commodity hardware .... as fate would have it, a couple years later he had to give one of the major m'soft ha presentations.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

osi archeological reference from ietf-pkix list

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: osi archeological reference from ietf-pkix list
Newsgroups: alt.folklore.computers
Date: Wed, 19 Feb 2003 19:23:03 GMT

osi archeological thread from ietf-pkix list:
https://www.garlic.com/~lynn/aadsm13.htm#21 A challenge
https://www.garlic.com/~lynn/aadsm13.htm#17 A challenge

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

difference between itanium and alpha

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: difference between itanium and alpha
Newsgroups: alt.folklore.computers
Date: Wed, 19 Feb 2003 20:32:08 GMT

J. Clarke writes:
If you don't want the OS that a particular vendor sells with a machine, then buy from a different vendor. <http://www.linux.org/vendor/system/desktop.html> lists a number of vendors who will sell you a machine with Linux preinstalled. In addition to those and Dell, HP sells their workstations with Linux if you want it.

... and there is walmart
http://www.pcworld.com/news/article/0,aid,109194,00.asp

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

COMTEN- IBM networking boxes

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: COMTEN- IBM networking boxes
Newsgroups: alt.folklore.computers
Date: Wed, 19 Feb 2003 22:14:23 GMT

Lars Poulsen writes:
I worked at ACC from 10/1980 to 01/1990. ACC was basically founded to build the channel attached 1822 interface for the mainframes of ARPA- net connected universities. By 1985, the network was transitioning from 1822 IMPs to X.25, and the WAN links were moving to routers while end nodes attached to the LAN. ACC built a front-end processor that looked to the mainframe channel as if it were the 1822 adapter. This allowed the use of the TCP/IP software that UCLA had developed for the 1822 box.

1980 or 1970? imps were supposed to disappear in 1983

in the early '80s, fsd had access to series/1s with all sorts of boards ... mainframe channel boards, numerous telco interfaces, etc.

FSD had been paying to keep 2701s alive because they had full-channel T1 support (telecommunication box form the '60s). In the early '80s, FSD funded the Zirpel card for the series/1 that support full-channel T1 .... as a replacement for all the old 2701s (remember the NSFNET1 T1 in the late '80s .... were actually 440kbit/sec cards in PC/RT with multiple 440kbit channels multiplexed into T1; there was a joke if you allowed that channels were flowing thru telco multiplexed channels, then the network could be considered at least T3/44mbit and in some places T5 ... aka a least somewhere in the telco network one of the nsfnet 440kbit channels might be multiplexed in a T5 channel).

the ncp rewrite for s/1 done by baby bell in the early '80s also used mainframe channel card for the s/1 ... where they emulated 37xx but faked out the mainframe vtam to believe all resources were cross-domain (every mainframe vtam was effectively told that the resources were owned by some other vtam .... when, in fact, the resources were "owned" by the distributed s/1s). from previous post in this thread:
https://www.garlic.com/~lynn/2003c.html#76 COMTEN- IBM networking boxes

in the early to mid '80s for HSDT .... for every T1 link that we were using HYPERchannel A7xx boxes on .... we had a Series/1 with Zirpel card sitting next to ... and be able to compare the HYPERchannel thruput to the S/1 Zirpel thruput (on T1 links). We started out using clear channel T1s ... and then somewhere along the way the phone company wanted to discontinue the clear channel T1s and only provide us with channelized T1s (among other things loosing every 193rd bit). random hsdt refs:
https://www.garlic.com/~lynn/subnetwork.html#hsdt

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

OT: Attaining Perfection

Refed: **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: OT: Attaining Perfection
Newsgroups: comp.arch
Date: Wed, 19 Feb 2003 22:52:30 GMT

Paul Wallich writes:
"The only thing that costs more than adding weight to an aircraft deisgn is taking it out." -- Norm Augustine (former CEO, Lockheed etcetera)

boyd's attempt at taking all the weight out of the f15 .... and when he was done they started putting it back in ... and they did something similar to the f16.

there is the corollary ... KISS

one reference ... although his biographies go into much more detail
http://www.codeonemagazine.com/archives/1997/articles/jul_97/july2a_97.html

from above:
The study showed that it was possible to have a fighter that was twice as maneuverable with twice the mission radius of the F-4D Phantom while weighing only 17,050 pounds. The study generated enough interest and gained enough attention to be a candidate for then Deputy Secretary of Defense David Packard's directed "experimental prototype" program. The rest is history: an airplane that has made history.

Boyd's theories are now used as a yardstick for measuring and comparing the maneuvering capability of all modern fighters. He planted the seeds for the Lightweight Fighter, cultivated those seeds, and helped harvest them in 1972 while playing a major role in defining the requirements that were put into the request for proposal for the Lightweight Fighter. The proposal led directly to the YF-16 and YF-17 technology demonstration prototypes. His influences can be seen in such world-class fighters as the F-15, F-16, and F/A-18.

... other refs:
https://www.garlic.com/~lynn/subboyd.html#boyd

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

CA-RAMIS

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: CA-RAMIS
Newsgroups: bit.listserv.ibm-main,alt.folklore.computers
Date: Thu, 20 Feb 2003 00:33:17 GMT

"edward konig" writes:
Right. The FOCUS people used to work for RAMIS, then split off 30+ years ago to form their own successful take-off.

http://www.decosta.com/Nomad/tales/history.html

note that in the above ... NCSS was some people from cambridge science center (and at least one person that worked on cp/67 from lincoln labs) going off and forming a cp/67 time-sharing service bureau

also in above note that NCSS began developing a alternative and released it under name of NOMAD ... while FOCUS was then released on Tymshare's (vm/370 .... a follow on to cp/67) time-sharing service.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Card Columns

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Card Columns
Newsgroups: alt.folklore.computers,comp.lang.pl1
Date: Thu, 20 Feb 2003 14:43:43 GMT

"Pointless Harlows" writes:
It must have been 1973 - mini-skirts were particularly short that year.

sometime in the early to mid '70s those student tables in classrooms got the panel across the front .... i remember hearing them referred to as modesty panels.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

CA-RAMIS

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: CA-RAMIS
Newsgroups: bit.listserv.ibm-main,alt.folklore.computers
Date: Sat, 22 Feb 2003 11:29:30 GMT

Shawn B. Wikle writes:
Ahhhhh. Very interesting! Thanks for the reference!

Shawn

well then you might also find interesting the recent SQL question from comp.database.theory (with a little compare and swap instruction drift):
https://www.garlic.com/~lynn/2003c.html#75 The relational model and relational algebra - why did SQL become the industry standard?
https://www.garlic.com/~lynn/2003c.html#78 The relational model and relational algebra - why did SQL become the industry standard?

note that ramis, nomad, focus, and sql all started out on vm systems (aka SQL started with system/r on vm at san jose research ... before tech. transfer to endicott for sql/ds ... with later tech transfer from endicott back to stl for db2).

slightly rounding out ref. ... both ncss and tymshare were cp/cms timeshare services. also started slightly after ncss ... another cp/cms timshare service was IDC.

and of course the largest such ... bigger than tymshare, ncss, and idc combined was HONE ... the internal online support for all the branch office and field people in the world (which started out with cp/67 and then moved to vm/370):
https://www.garlic.com/~lynn/subtopic.html#hone

one of hone problems was that there was frequent executive churn/promotions in DPD. A new DPD executive would come in that had been taught that MVS solved all customer problems and eventually realize that HONE was built on cp/cms platform. He would then ask it be moved to MVS platforms. After extended period time, it would eventually be shown again that wasn't practical and the whole thing would be quitely dropped. For an extended number of years .... possibly half of HONE development resources went into repeatedly having yet another demonstration that it wasn't practical to host HONE on MVS.

so for even more archeological drift, recent thread on PCM & COMTEN:
https://www.garlic.com/~lynn/2003c.html#70 COMTEN- IBM networking boxes
https://www.garlic.com/~lynn/2003c.html#76 COMTEN- IBM networking boxes
https://www.garlic.com/~lynn/2003c.html#77 COMTEN- IBM networking boxes
https://www.garlic.com/~lynn/2003c.html#79 COMTEN- IBM networking boxes
https://www.garlic.com/~lynn/2003d.html#13 COMTEN- IBM networking boxes

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Efficent Digital Signature Schemes

Refed: **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Efficent Digital Signature Schemes.....
Newsgroups: sci.crypt
Date: Sat, 22 Feb 2003 11:59:03 GMT

Paul Crowley writes:
Is it running on a smart card?

It sounds like DSA-512 would be an exact fit. The group parameters take 642+20=148 bytes of storage in total, and the private key a further 20 bytes. You can safely share the group parameters between all users if you have more storage in a shared ROM than in non-volatile storage. You can also use "precomputation" for very fast signing. See "Applied Cryptography" for details.

note aads chip strawman
https://www.garlic.com/~lynn/x959.html#aads

is hardware token that does fips186-2/x9.62/ecdsa signatures

disclaimer ... this is the AADS chip strawman that i've been going on about for some number of years.
https://www.garlic.com/~lynn/x959.html#aads

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

PC history, was PDP10 and RISC

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: PC history, was PDP10 and RISC
Newsgroups: alt.folklore.computers,alt.sys.pdp10
Date: Sat, 22 Feb 2003 16:42:38 GMT

johnl@iecc.com (John R. Levine) writes:
If Boca was Siberia, IBM Instruments was on the moon, a little company that IBM bought to plug a tiny hole in the product line. The S9000 was a lab computer that only coincidentally happened to be sort of usable as a desktop PC. It was way too expensive to sell in the desktop market and competed with PDP-11s, not Apples.

ibu & acorn
https://www.garlic.com/~lynn/2003d.html#9 IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
https://www.garlic.com/~lynn/2002g.html#79 Coulda, Woulda, Shoudda moments?

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

COMTEN- IBM networking boxes

Refed: **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: COMTEN- IBM networking boxes
Newsgroups: alt.folklore.computers
Date: Sun, 23 Feb 2003 17:41:12 GMT

Anne & Lynn Wheeler writes:

https://www.garlic.com/~lynn/2003c.html#76 COMTEN- IBM networking boxes

and of course that was a brain check:
https://www.garlic.com/~lynn/2003c.html#76 COMTEN- IBM networking boxes

and of course the refernece to "ncp" in that post
https://www.garlic.com/~lynn/2003d.html#13 COMTEN- IBM networking boxes

was with respect to ncp/pu4 that ran in 37xx boxes (basically sna's imp/fep) and not to arpanet's host "ncp" implementation. misc. arpanet host "ncp" refs:
https://www.garlic.com/~lynn/internet.htm#26 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/2000.html#67 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/internet.htm#27 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/2000.html#72 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/internet.htm#28 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/2000.html#73 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/internet.htm#29 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/2000.html#74 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/internet.htm#31 Difference between NCP and TCP/IP protocols
https://www.garlic.com/~lynn/2000.html#85 Difference between NCP and TCP/IP protocols

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

PDP10 and RISC

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: PDP10 and RISC
Newsgroups: alt.folklore.computers,alt.sys.pdp10
Date: Sun, 23 Feb 2003 20:37:57 GMT

Brian Inglis writes:
The 8KB page size would have been about right to allow growth for total system memory on the VAX, and would probably have made emulating 11 mode more straightforward. 2KB pages were used in small IBM 370 OSes with up to 16MB system memory, and 2KB was too small for full memory systems; 4KB was used in bigger IBM 370 OSes and was okay for 16MB up to 32MB system memory; 8KB would have been about right for the VAX time frame and target system memory sizes, and was eventually used in the Alpha.

2kb was probably 512kbyte or less. VS1 (which had 2k pages) running under VM with ecps and handshaking ... typically ran faster on 1mbyte (or larger real machine) under VM than running standalone (even after taking out VM's fixed kernel requirements of up to 200k ... aka 20 percent of one mbyte real machine).

In handshaking mode, VS1 was given a 16mbyte virtual address space by VM (VS1 treated it as if running on 16mbyte real machine) ... and VS1 didn't do any paging ... but let VM handle it all using 4k pages ... aka VS1 normally had a single 16mbyte virtual address space ... and in handshaking mode it was mapped one-to-one with the virtual machine 16mbyte address apace.

later in 3081 time frame with 3380s (early '80s) ... "big" pages were introduced for both VM and MVS. The hardware translation was still all 4k pages, however "big" pages was change to both vm & mvs page I/O to do transfers in units of full-track ... aka 40k or ten pages at a time. when being replaced .... ten pages from a virtual address space were collected together and written out in one i/o. On a page fault for any of the ten 4k pages in a unit ... all ten pages would be read in one operation.

the issue that dominated wasn't so much the packing in real storage but the amount transferred in a single unit ... given the i/o operational characteristics. by the time-frame of 1mbyte real storages, I/O was more of a thruput constraint than real storage ... and strategies (like caching) were starting to appear that attempted to leverage real storage for i/o bottleneck compensation.

past discussions of 360/67 w/105 4k pages (after fixed kernel requirements) and 3081k with >7000 4k pages (after fixed kernel requirements):
https://www.garlic.com/~lynn/93.html#31 Big I/O or Kicking the Mainframe out the Door
https://www.garlic.com/~lynn/95.html#10 Virtual Memory (A return to the past?)
https://www.garlic.com/~lynn/98.html#46 The god old days(???)
https://www.garlic.com/~lynn/99.html#4 IBM S/360
https://www.garlic.com/~lynn/2001f.html#62 any 70's era supercomputers that ran as slow as today's supercomputers?
https://www.garlic.com/~lynn/2001l.html#40 MVS History (all parts)
https://www.garlic.com/~lynn/2001l.html#61 MVS History (all parts)
https://www.garlic.com/~lynn/2001m.html#23 Smallest Storage Capacity Hard Disk?
https://www.garlic.com/~lynn/2002.html#5 index searching
https://www.garlic.com/~lynn/2002b.html#11 Microcode? (& index searching)
https://www.garlic.com/~lynn/2002b.html#20 index searching
https://www.garlic.com/~lynn/2002e.html#8 What are some impressive page rates?
https://www.garlic.com/~lynn/2002e.html#9 What are some impressive page rates?
https://www.garlic.com/~lynn/2002i.html#16 AS/400 and MVS - clarification please

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Which Editor

Refed: **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Which Editor
Newsgroups: alt.folklore.computers
Date: Tue, 25 Feb 2003 03:08:09 GMT

Joe Morris writes:
XEDIT as in IBM's VM product, ported almost completely to the PC by Mansfield Software Group (www.kedit.com).

xedit in large part xavier delamberterie responsible in the late '70s with various help from people in endicott, csc, and yorktown. there were some issues at the time that xedit was somewhat traditional development project while there were several editors (nost notably red) available internally that had gone thru numerous enhancement iterations with feedback based on possibly tens of thousands of internal users. it wasn't so much that xedit was a bad editor ... it was that red had significantly more internal development effort based on real life use put to it (somewhat akin to current open source lore).

red, ned, xedit, and edgar had similar capability ... but simple cpu usage test that i did (summery of '79) of the same set of operations on the same file by all editors showed the following cpu uses:


RED        2.91/3.12
EDIT       2.53/2.81
NED       15.70/16.52
XEDIT     14.05/14.88
EDGAR      5.96/6.45
SPF        6.66/7.52
ZED        5.83/6.52

"edit" is the standard vanilla cms editor from the time. xedit eventually gained rex(x) scripting capability.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

CPU Impact of degraded I/O

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: CPU Impact of degraded I/O
Newsgroups: bit.listserv.ibm-main
Date: Tue, 25 Feb 2003 03:25:34 GMT

Patrick_Falcone@AON.COM (Patrick Falcone) writes:
I would agree with what Matt said but add that it can depend on the profile of the workload you are helping. If for instance you batch work is say 85% I/O and 15% CPU then you would probably see some additional CPU. If it's a heavily weighted CPU type workload then you may not view any significance. I would tend to think that most common batch workloads would weigh more heavily on I/O versus CPU so you may see some additional CPU.

For interactive workloads it becomes a measure of productivity. If an End User is use to 2 - 5 second response time and all of a sudden you reduce this time to 1 second or less the Users most likely will become more productive which translates to more transactions which will increase CPU.

Since I/O is the slowest of the three (CPU, Storage, I/O) I would tend to think that increasing the speed of I/O delivery would increase the use of CPU but I'll bow to the heavyweights for their comments. Of course relieving this bottleneck could cause you to find you have another potential bottleneck.

the issue is that CPU utlization can increase with effective i/o thruput; more work being done in less time. I believe that the question was whether or not that degraded I/O increased the cpu overhead for handling I/O (aka the avg. overhead pathlength per I/O operation increases).

long ago and far away ... in STL ... there was a situation where a bunch of "local" 3274s were removed from direct channel attach and put at the end of HYPERchannel channel extender (300 people working on ims moved to remote location). The overall thruput of the system went up 10-15% and the avg. system response decreased (improved) ... with no noticeable degradation by the 300 "remoted" individuals.

The issue was that the local HYPERchannel boxes that sat directly on the channels had significantly lower channel busy per operation than than 3274s. The same operations could be done with much less channel overhead busy ... which made additional channel capacity available for lots of other operations .... increasing overall system thuput.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

CPU Impact of degraded I/O

Refed: **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: CPU Impact of degraded I/O
Newsgroups: bit.listserv.ibm-main
Date: Tue, 25 Feb 2003 03:38:43 GMT

Anne & Lynn Wheeler writes:
long ago and far away ... in STL ... there was a situation where a bunch of "local" 3274s were removed from direct channel attach and put at the end of HYPERchannel channel extender (300 people working on ims moved to remote location). The overall thruput of the system went up

the 300 had subsecond respond for nominal interactive operations before the move and it didn't measurably change after the moved of 300 people/terminals to remote site (however, all of the more complex things improved for everybody with the overall system thruput increase of 10-15 percent).

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Which Editor

Refed: **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Which Editor
Newsgroups: alt.folklore.computers
Date: Tue, 25 Feb 2003 12:37:26 GMT

Brian Inglis writes:
The difference is probably that Xedit was documented and released to customers and came with extensible scripting capability (from VM/SP and EXEC2): think emacs-like in CMS. SPF, edit and edgar seemed to be pretty basic editors with little extensibility. Never heard of ned and red: maybe internal only; like so much that was good on VM and CMS and never saw the light of FCS?

the issue was all before xedit was released to customers as a product. the discussion was whether or not endicott would be better off continuing with the development of xedit .... or better off picking up one of the other internal (more mature) editors (like red) as the base for doing further development and releasing as a product.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Antiquity of Byte-Word addressing?

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Antiquity of Byte-Word addressing?
Newsgroups: comp.arch,alt.folklore.computers
Date: Tue, 25 Feb 2003 23:21:13 GMT

"Glen Herrmannsfeldt" writes:
Two megabytes was big for a 360. Those bytes are all magnetic cores with wires through them! Most models didn't have 24 bits of physical address, and no virtual memory yet.

I think the high end 370's (before the 303x series) had a maximum of 8MB physical memory, though 16MB virtual, and now monolithic memory.

360/67 had virtual memory ... and supported both 24bit virtual and 32bit virtual addressing. you could get up to 1mbyte of real storage. in a duplex system ... each processor having 1mbyte of real storage for 2mbyte of total addressable real storage.

high end 168s you could deck out with 16mbytes of real storage.

3033 was having some interesting issues and had a gimmick to support up to 64mbytes of real storage with only 24bit addressing. the PTE (page table entry) in 4k-page mode had two unused/undefined bits ... out of the 16 bits (12 bit real 4k page number ... for 24bit real storage addressing, PTE invalid flag bit, page protection flag bit and two unused bits). The 3033 redefined the two unused/undefined bits to be allowed as real 4k page number ... allowing 14bit real 4k page number for 26bit real address or 64mbytes.

CCWs had to be located in the first 16mbytes of real storage ... but could point to IDAL which could specify 31bit addresses.

prior discussion of 3033 26bit real addressing
https://www.garlic.com/~lynn/2002c.html#40 using >=4GB of memory on a 32-bit processor

3081, 370-xa introduced 31bit virtual addressing.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

[urgent] which OSI layer is SSL located?

Refed: **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: [urgent] which OSI layer is SSL located?
Newsgroups: comp.security.misc
Date: Wed, 26 Feb 2003 12:43:01 GMT

zhongmeiyi@yahoo.com.sg (megan) writes:
no, this isn't a homework question.

Some have said that TLS, being Transport Layer Security would mean that SSL is belongs to the tranport layer, but others have said that SSL is a Session layer protocol. There are also people saying its both layers

total aside ... OSI defines both the 7-layers for moving data as well as "service" functions (which sits beside/outside the 7-layers and can span layers). OSI is somewhat viewed as effort to define infrastructure for point-to-point, low-speed, high error rate copper.

For instance LANs violate OSI ... basically covering layers 1, 2, and part of 3. Various parts of international (ISO) and national (ANSI) organizations claimed that they couldn't consider standards that violated OSI. This got them somewhat schizo when LANs got accepted as ISO standard via IEEE. An example is x3s3.3 (ANSI responsible for standards for layers 3 & 4) couldn't consider standards that interfaced to LAN because they violated ISO guidelines that prohibited standards that didn't conform to OSI.

SSL is something of anomoly ... it was transport layer protocol that was implemented in an application. Note also that IP (internetworking) also violates OSI since it sits in a non-existent place (in the OSI model) between transport and networking.

misc. past discussions:
https://www.garlic.com/~lynn/subnetwork.html#xtphsp

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

Refed: **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: comp.arch,alt.folklore.computers
Date: Wed, 26 Feb 2003 14:43:33 GMT

jmaynard@thebrain.conmicro.cx (Jay Maynard) writes:
There's a problem with OS/360 MVT TSO: it blows up rather badly if it runs above 8 MB. I've asked several folks at SHARE this week whowere around then, and basically got the answer that they didn't know of any 360s that had more than 8 MB on them, and damned few of those. I don't think it was a conscious decision so much as a general feeling that 24 bits would address more memory than anyone thought would ever be economical to hang on a machine - a cultural thing rather than an explicit decision.

some number of machines had 8mbytes of (ampex, etc) LCS installed ... in addition to their standard memory (resulting in >8mbytes). also has been mentioned, 360/67 had both 24bit virtual and 32bit virtual addressing.

misc. ampex/lcs memory
https://www.garlic.com/~lynn/2000.html#7 "OEM"?
https://www.garlic.com/~lynn/2000e.html#2 Ridiculous
https://www.garlic.com/~lynn/2000e.html#3 Ridiculous
https://www.garlic.com/~lynn/2001f.html#51 Logo (was Re: 5-player Spacewar?)
https://www.garlic.com/~lynn/2001j.html#15 Parity - why even or odd (was Re: Load Locked (was: IA64 running out of steam))
https://www.garlic.com/~lynn/2003c.html#62 Re : OT: One for the historians - 360/91
https://www.garlic.com/~lynn/2003c.html#63 Re : OT: One for the historians - 360/91

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

SSL questions

Refed: **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: SSL questions
Newsgroups: sci.crypt
Date: Thu, 27 Feb 2003 21:46:48 GMT

PunkroyREMOVETHIS@DrQue.net (Punkroy) writes:
So that I am clear in what many seem to be saying: If anyone compromises the server certificate, all communications that used that certificate CAN be decoded. Is that correct? If so, am I the only one who feel that is a rather large open door for all those e-commerce sites. A disgruntled employ gets access to the server certificate and could decrypt any session that certificate was used for? Okay, but the certificate still has a password on the private key, right? But I've read server certificate can be stored without the private key encrypted. In the setup I did, I read about this option so you didn't have to enter a password when the webserver first started. Anyone have an idea of how many admins don't encrypt the certificate private key?

For some reason, I had always thought the certificate was used only for authentication and that both the server and client would generate a temporary key set used to encrypt traffic only for that session. After the session was complete, the key sets were discarded and the data they sent back can not be decrypted by anyone-- including the original sender and receiver (less bruteforce attack on session key, factoring public key, ect). If that is not the case, why? I thought about the possibility of speed being an issue-- but the client has to generate a keyset. True, the server might be doing hundreds of connections at once, but is the trade off really worth the risk? Is there a system like the one I've outlined above already in SSL? If so, please give me a link-- that is what I am interested in using.

Since I am on the subject of server certificates, I might as well ask this question as well: Does the root authority who signs a server certificate ever get the private key? It seems this question should also be a "no" answer, but after learning more about certificates from this thread, I'm not sure I trust anything done by SSL.

if you know somebody's public key .... you can encrypt stuff with that public key and only the entity with the corresponding private key can decrypt it. various protocols generate a random secret key ... encrypt the secret key with the public key and then encrypt the actual data with the secret key. only the entity with the corresponding private key can recover the secret key and then decrypt the rest of the message.

the certificate isn't held under password look & key ... the private key is suppose to be held in some sort of protective control. a certificate is some digitally signed information that effectively creates a trusted binding between some pieces of information and a public key.

for ssl server certificates there is the complex chain of processes to 1) provide for privacy of data (aka encryption) and 2) validate that the entity that you think you are communicating with is really that entity.

browsers have tables of trusted (CA, certification authority) public keys. CAs are responsible for validating associations between domain names and public keys ... and creating digital signed certificates that assert the binding between a domain name and a public key. the integrity of the certificate is validated by using one of the CA public keys to confirm the CA's digital signature (on certificates).

browsers that have validated the integrity of the certificate also check that the URL that somebody typed in has a domain name that matches that in the certifcate ... that establishes the equivalence between the domain name and a public key.

the browser generates a random secret key, encrypts some data with the random secret key, encryptes the random secret key with the public key. given
1) the validity of a certificate, 2) the equivalence between the typed in URL and the domain name in the certificate 3) the certificate's domain name and the certificate's public public key, 4) the encryption of the random secret key with the public key 5) the encryption of the data with the random secret key

then the assumption is that only the web server in the original typed in URL can decrypt the data (since only that web server has the private key that can recover the random secret key).

a big reason for all of this infrastructure is a concern regarding the integrity of the domain name infrastructure's ability to provide trusted serving up of domain name to ip-address mapping (aka a compromise of the domain name infrastructure so that traffic is routed to a fraudulent ip-address).

a problem is that the authoritative agency for domain names is the domain name infrastructure .... nominally certification authorities aren't the actual agency responsible for the information they are certifying .... they must corroborate the information in a certificate application with the authoritative agency ... in the case of domain names, is the domain name infrastructure ... aka the certification authorities (for ssl server domain name certificates) are dependent on the very same infrastructure that everybody has integrity concerns about.

so there have been some suggestions (somewhat prompted by certification authorities) about how to improve the integrity (vulnerabilities) of the domain name infrastructure. a catch-22 is that if the integrity of the domain name infrastructure is improved for the certification authorities ... it is also improved for everybody (somewhat mitigating the concern about its integrity and the original justification for having ssl certificates).

lots of past discussion:
https://www.garlic.com/~lynn/subpubkey.html#sslcerts

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

SSL questions

Refed: **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: SSL questions
Newsgroups: sci.crypt
Date: Thu, 27 Feb 2003 22:01:18 GMT

PunkroyREMOVETHIS@DrQue.net (Punkroy) writes:
If so, am I the only one who feel that is a rather large open door for all those e-commerce sites. A disgruntled employ gets access to the server certificate and could decrypt any session that certificate was used for?

note that the certificate isn't able to decrypt any SSL data ... the certificate is suppose to establish the equivalence between someinformation and a public key (i.e. domain name .... information digitally signed by that web server's private key can be validated by the public key in the corresponding certificate, information encrypted with the web server's public key can only be decrypted with the web server's private key).

SSL only encrypts the data in flight. I'm not actually aware of a compromise of a credit card number transmitted in the clear on the internet. however, the business processes associated with credit cards require lots of uses of credit card numbers in the unencrypted form. large-scale harvesting of these numbers (at rest) has been established (as opposed to any harvesting of numbers in flight) .. some large scale harvesting mention:
https://www.garlic.com/~lynn/subintegrity.html#fraud

some discussion of e-commerce:
https://www.garlic.com/~lynn/aadsm5.htm#asrn2
https://www.garlic.com/~lynn/aadsm5.htm#asrn3
https://www.garlic.com/~lynn/2001i.html#52

and somewhat related assurance:
https://www.garlic.com/~lynn/subintegrity.html#assurance

standard that eliminates PAN/account number as point of fraud:
https://www.garlic.com/~lynn/x959.html#x959
https://www.garlic.com/~lynn/subpubkey.html#privacy

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: comp.arch,alt.folklore.computers
Date: Fri, 28 Feb 2003 01:11:41 GMT

hancock4@bbs.cpcn.com (Jeff nor Lisa) writes:
Indeed, on the original IBM PC, wasn't the full 640K an extra cost option--that is, the default memory for the machine was much smaller and cost less?

In other words, in the early days of PCs, while memory was still cheaper than core memory, there were still cost limitations. Indeed, wasn't the original DOS 2.0 fairly limited to even handling "extended" or "expanded" memory on the 8086 machine? (I never could understand the differences between those two terms).

memory definition:
https://www.garlic.com/~lynn/2001f.html#37 Ancient computer humor - Memory
https://www.garlic.com/~lynn/2001k.html#73 Expanded Storage?

standard entry consumer machine had 64k ... more memory cost more money.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

Refed: **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
 monopoly
Newsgroups: alt.folklore.computers
Date: Fri, 28 Feb 2003 01:30:31 GMT

lars@bearnip.com (Lars Duening) writes:
Meaning that all the nice humid air from the pacific can't get to us because of the mountains, and the nice humid air from the atlantic runs out of steam somewhere in the vicinity of Kansas.

backside of the mountains tend to be semi-arid desert ... backside of the sierras, backside of the cascades, backside of the rockies, even the backside of the santa cruz mountains.

backside of the cascades has the columbia river ... the grand coulee dam not only is one of the largest hydro electric dams ... but also pumps water into the coulee which then flows into large irrigation system serving the semi-arid desert along the backside of the cascades ... currently something like half million acres.

past ref:
https://www.garlic.com/~lynn/2002n.html#43 VR vs. Portable Computing

misc other
http://users.owt.com/chubbard/gcdam/html/photos/irrigation.html
http://users.owt.com/chubbard/gcdam/
https://web.archive.org/web/20090513013752/http://www.usbr.gov/power/data/sites/grandcou/grandcou.html
http://www.ccrh.org/comm/moses/moseslake.html

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: comp.arch,alt.folklore.computers
Date: Fri, 28 Feb 2003 12:35:03 GMT

shoppa@trailing-edge.com (Tim Shoppa) writes:
Just for comparison: a 33 MHz 80386 - or for that matter a late 70's VAX 11/780 @ nominal 1 MIPS - can easily saturate a 10 Mbit Ethernet with a FTP or Web server. That's six times the bandwidth of a full T-1.

I was ridiculed a few months ago for suggesting that a 800 MHz server was more than enough for serving a 384kbit SDSL connection.

for the mainframe tcp/ip product, i did the rfc 1044 support ... and testing at cray research saturated the 1mbyte/sec channel on 4341-clone using nominal percentage of the cpu. by comparison the regular support going thru 8232 could saturate a 3090 processor at 44kbytes/sec.

there is a big issue of the efficiency of the implementation as well as the efficiency of the protocol. nominal FTP activity sets up a long running TCP session and transfers lots of data at maximum MTU. worst case HTTP is single minimal sized packet requiring 7 packet exchange for tcp setup/teardown. however, in principle, 800MHz anything should be overkill for 384kbit.

somewhat indirect comparison since it is from (old) linpack table for floating point


Computer                       N=100(Mflops)
--------------------------------------------
Gateway 2000 G6-200 PentiumPro            62
IBM 3090/180J                             10
AMD 486DX5-133                           4.4
DATEK 80386-33 /w 64KB Cache             .27
IBM 4341 MG10                            .19
VAX 11/780 FPA                           .14
Northgate 386/387 (25MHz)                .11
VAX 11/780 FPA                           .11

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

YKYBHTLW

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: YKYBHTLW...
Newsgroups: alt.folklore.computers
Date: Fri, 28 Feb 2003 14:47:35 GMT

Giles Todd writes:
I didn't realize that PIN-based payments (as opposed to PIN-based ATM withdrawals) had caught on in North America. They are certainly a novelty in the UK, although nothing new in 'Old Europe'.

some issue in the walmart/retail legal action is offline (i.e. non-PIN) debit vis-a-vis online (pin) debit (in part because offline debit fees are more in line with credit fees).

see debit card lawsuit at
http://www.cardweb.com/cardtrak/pastissues/apr02.html

some topic drift with respect to PANs:
https://www.garlic.com/~lynn/2003d.html#30 SSL questions

standard for all payments (debit, credit, stored-value, echeck, etc) in all environments (POS, face-to-face, e-commerce, etc):
https://www.garlic.com/~lynn/x959.html#x959

digital signature token that could be used in above:
https://www.garlic.com/~lynn/x959.html#aads

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: comp.arch,alt.folklore.computers
Date: Fri, 28 Feb 2003 15:19:35 GMT

"Rupert Pigott" writes:
... I wasn't aware that FP was particularly relevant to TCP/IP stacks ... :)

as per statement it was somewhat indirect. that was the first table that i could find that had comparison involving 386, 780s, 4341s, 3090s, and pentiums (although only going up thru 200mhz) ... since they all compared FP ... there was possibly some correlation between the relative difference between the 5-6 machines in other areas ... aka the relative FP performance might also be some indication of the relative non-FP performance (between the various machines).

there was no statement with regard to how many FPs were required per TCP op .... there was statements with regard to 4341 implementation not being processor bound (at 1mbyte/sec) and 3090 being processor bound (at 44kbyte/sec) and the relative (FP) thruput of a 4341 processor to a 3090 processor being possibly factor of 50 (and a 780 being somewhat in the same class as 4341).

The 4341 and 3090 implementation was the same .... except differing by support I had done for rfc1044 vis-a-vis the standard 8232 support (in effect the same tcp/ip stack/implementation) showed something like a two orders of magnitude processor-use difference (or more) between the rfc1044 support and the 8232 support and a factor of 24 difference in data thruput ... approaching four orders of magnitude (aka 10,000 times) difference in instructions executed per byte moved).

ref:
https://www.garlic.com/~lynn/2003d.html#33 Why only 24 bits on S/360?

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: comp.arch,alt.folklore.computers
Date: Fri, 28 Feb 2003 15:50:27 GMT

Robin KAY writes:
If your router provides a firewall then why do need an additional firewall on the machine? Mine is set to block all incoming traffic.

some number of the personal firewalls will block by application (both incoming and outgoing) ... where standard packet-filtering firewalls tend to be limited to port &/or ip addresses (in and out). for instance i don't mind if my browser goes out to a port 80 ... but i've periodically caught some spyware trying to get out to a port 80.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: comp.arch,alt.folklore.computers
Date: Fri, 28 Feb 2003 16:52:16 GMT

"Rupert Pigott" writes:
That snippet tells me something useful. Can't find any refs to the estimated instructions/byte moved by cdrom.com's FreeBSD FTP server running on a PPro... I'm sure I've seen that somewhere. I'm not finding the magic combo with Google, sigh.

IIRC it was a surprisingly small number. :)

note that wasn't the number of instructions per tcp operation ... it was that there was four order of magnitude difference (ten thousand times) in the number of instructions per tcp byte moved (i.e. if 1044 support executed say 5 instructions per byte ... then the 8232 support was executing nearly 50,000 instructions per byte).

the "fast" router pathlength presentation at 88(89?) IETF for gigabit link was around 120 instructions (with no data copies) and later got it into 85 instruction range. i think that standard 4.3 reno/tahoe tcp stack was around 5000 instructions and five data copies in the late '80s.

with regard:
https://www.garlic.com/~lynn/2003d.html#33 Why only 24 bits on S/360?
https://www.garlic.com/~lynn/2003d.html#35 Why only 24 bits on S/360?

somewhat related when getting started on ha/cmp,
https://www.garlic.com/~lynn/subtopic.html#hacmp

we wanted to know how much of a processor was being used in various tcp/ip operations. it was somewhat difficult to get really good total cpu use on unix ... so used some stuff that had done long ago and far away for capacity planning ... a lot of work evolving performance tuning into capacity planning went on in the late '60s and early '70s at the cambridge science center:
https://www.garlic.com/~lynn/subtopic.html#545tech
https://www.garlic.com/~lynn/submain.html#bench

so took a program that had a tight loop that ran for a known number of minutes.

after the loop program started, a tcp/ip application was started and transferred a fixed amount of data. it was run transferring a single byte as well as run transferring a couple hundred megabytes over enet. The difference in elapsed time for the loop program was assumed to be the total cpu use needed by tcp/ip application (and supporting infrastructure).

this information was used in various capacity planning calculations for the rs/6000 thruput for fddi, SLA (serial link adapter ... 220mbit/sec connection), hippi, and fcs. random ref:
https://www.garlic.com/~lynn/95.html#13

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

The PDP-1 - games machine?

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: The PDP-1 - games machine?
Newsgroups: alt.folklore.computers
Date: Sat, 01 Mar 2003 14:38:11 GMT

"Ross Simpson" <yeah_whatever> writes:
I've got a Time Life Book which talks about Computer Software. In this it talks about the PDP-1 computer from the early 1960s & about one of the games written for it called "SpaceWar", I was suprised to see that this game had many varients (done by various hackers at MIT).

cambridge science center had a 2250-4 (aka 2250 with 1130 as controller). somebody in the late '60s ported spacewar to the 2250-4. the 2250 keyboard (looked much like selectric keyboard) was split in half ... with key controls for two players on the left and right half. i don't know if the person that had done the 2250-4/1130 port was involved in the pdp-1 at all (although i don't believe it was a very large community ... since just about everybody at csc had been around ctss and/or other aspects of mit computing for some period).

I would sometimes bring my kids in on weekends and they would play while i worked ... they would also sometimes chase each other up and down the hall ... which would get complaints from other possible people also in on weekends.

picture of 2250-4
http://www.shubs.net/1130/functional/DisplayUnit.html

other 2250-4/1130 ref:
http://www.forth.com/Content/History/History1c.htm
http://ibm1130.org/lib

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
 monopoly
Newsgroups: alt.folklore.computers
Date: Sat, 01 Mar 2003 14:44:25 GMT

jmfbahciv writes:
Not if you have to do the work. Irrigation uses gravity to do the work; sprinkling requires a pump which requires electricity, not to mention a hell of a lot of pipe that needs to be maintained.

or a tractor with a drive belt connected to the pump. the thing that got me with moving pipes ... you would accumulate 5-10 lbs of mud on each boot ... and there was this sucking pressure just moving each foot. now days you fly over these fields where sprinkling is down in these large circles.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Authentification vs Encryption in a system to system interface

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Authentification vs Encryption in a system to system interface
Newsgroups: comp.security.misc
Date: Sat, 01 Mar 2003 15:06:30 GMT

"Edward A. Feustel" writes:
The authentication can be done mutually when the SSL channel is established. Apache (with the proper modification) permits the authenticated identity to be passed in an environment variable to the "application". The application can do what it will with that information. Whether the application can masquerade as the user in order to use the ACL system depends on the system, but the authenticated ID can be used to determine what the user can do within the file system by using the standard access control mechanisms that Apache provides.

in SSL with SSL domain name server certificates ... the real authentication is done when the CA is creating the certificate. the browser just checks that the URL typed in just matches the domain name in the certificate (is the browser really talking to the server that it thinks it is talking to). the vulnerability being addressed is possible integrity problems with the domain name infrastructure.

the catch-22 of course is that the CA is a certification operation, and must verify with the authoritative agency the validaty of the information being certified (i.e. is the entity requesting a certificate really the owner of that domain?). The authoritative agency for domain names is the domain name infrastructure ... so CAs are relying on the very same agency that have the original integrity issue giving rise to the certificate requirement.

so CAs have proposals for improving the integrity of the domain name infrastructure ... so that they can rely on the validity of the information ... but improving the integrity of the domain name infrastructure contributes to mitigating the requirement for having certificates in the first place.

similar explanation from sci.crypt:
https://www.garlic.com/~lynn/2003d.html#29 SSL questions
https://www.garlic.com/~lynn/2003d.html#30 SSL questions

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: alt.folklore.computers
Date: Sat, 01 Mar 2003 15:16:33 GMT

Morten Reistad writes:
We were running an ISP-mix of packets, where number of routes, packet distributions on source/destination ports and addresses, packet sizes and bursts were matched to statistical data from real ISP operations. Same packet stream that went into a lot of other benchmarks.

The average packet size was 272 octets gross, btw.

This was the business load for a number of organizations, and router offers were generally woefully inadequate in such performance. This benchmark was presented as a protest to cisco hype.

The cisco reps were never friends with me for years afterwards,

i believe there was an IETF presentation that packet size profile was strong bi-model distribution ... combination of max. MTU packets ... in combination with small packets.

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

Authentification vs Encryption in a system to system interface

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Authentification vs Encryption in a system to system interface
Newsgroups: comp.security.misc
Date: Sat, 01 Mar 2003 21:14:49 GMT

"Edward A. Feustel" writes:
While the above is true, the persons in the original article appeared to be discussing Client Authentication and Authorization rather than Server Authentication.

Ed Feustel

oops, well only one of the public keys are necessary for key exchange for encryption. one of the original "client" authentication implementations (we had sign-off ... so we could require that the commerce server did mutual authentication to the payment gateway ... even tho there wasn't anything specified at the time for mutual/client authentication):
https://www.garlic.com/~lynn/aadsm5.htm#asrn2
https://www.garlic.com/~lynn/aadsm5.htm#asrn3

one of the suggestions for webserver doing client authentication was provide a stub to radius ... and then radius administrative control could decide whether individual users/accounts were authenticated by password, chap, digital signature, etc.
https://www.garlic.com/~lynn/subpubkey.html#radius

aka administrator could specify to radius at individual account level what mode of authentication would be used. it would have the advantage for large operations that also inlcuded things like web hosting ... a single adminstrative infrastructure for authentication across the whole operation (regardless of the type of authentication).

--
Anne & Lynn Wheeler | lynn@garlic.com - https://www.garlic.com/~lynn/
Internet trivia, 20th anniv: https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

Refed: **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
monopoly
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Sun, 02 Mar 2003 14:56:04 GMT

bdc@world.std.com (Brian 'Jarai' Chase) writes:
Also. Does anyone else find it sort of ironic that Apple is counting on IBM's PowerPC 970 for use in their next generation systems? Business is business, but it's a long way from 1984.

just an extension of fort knox. pre 1984 .... fort knox was a project to migrate most of the sundry microprocessors used in controllers and various computers to 801/risc (rochester products, low & mide range 370s ... aka boeblingen & endicott, numerous controllers using uc.5s, jib-primes, etc).

fort knox got killed ... and some number of the people went off to various other places (including outside the company; slightly amd related; i believe 29k had at least one such person) to build risc processers.

ROMP was 16bit 801 for the OPD (office product division) displaywriter follow-on. when that got killed ... the group retargted the hardware for unix workstation ... getting the group that had done the AT&T port to the pc for pc/ix to do one ... resulting in pc/rt & aixv2.

folklore has it that after Future System (FS) got killed, some number of people went off to rochester and built it anyway ... resulting in the s/38. The s/38 follow-on, as/400 was initially built on cisc but was retargted to power/pc ... and so could be considered "fort knox" ... just delayed by something like 15 years.

misc 801 &/or fort knox references
https://www.garlic.com/~lynn/subtopic.html#801

misc FS references:
https://www.garlic.com/~lynn/submain.html#futuresys

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Why only 24 bits on S/360?

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Why only 24 bits on S/360?
Newsgroups: comp.arch,alt.folklore.computers
Date: Sun, 02 Mar 2003 15:03:29 GMT

rpw3@rpw3.org (Rob Warnock) writes:
It was probably Interdata, later bought by Perkin-Elmer. More info here: <URL:http://simh.trailing-edge.com/interdata.html>

in the 60s, when i was undergradudate, we used interdata/3 to build plug-compatible 360 telecommunication controller (reverse engineered channel interface, etc) ... and subsequently got blamed for originating the 360 PCM business. random bits:
https://www.garlic.com/~lynn/submain.html#360pcm

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
monopoly monopoly monopoly
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Sun, 02 Mar 2003 23:24:41 GMT

winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") writes:
Apple and IBM were also the joint parents of Taligent. (Whatever happened to Taligent, anyway?)

it seemed like it was picked up by java ... at least the bldg and people ... i believe some people continued in their old offices.

when we started ha/cmp ... the executive we reported to had come from motorola. he then headed up somerset .... and then went on to be president of mips.
https://www.garlic.com/~lynn/subtopic.html#hacmp

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: alt.folklore.computers,comp.os.vms
Date: Mon, 03 Mar 2003 14:15:18 GMT

jmfbahciv writes:
NOpe. You have to understand that these guys did not comment for the benefit of the code reader. The comments were for their benefit and ONLY their benefit. It wasn't until the company got too big that commentary was deemed necessary for the maintenance group. In the beginning, maintenance was done by the developers. I don't know how to communicate this to you people who 1. haven't done OS development as their primary business and 2. can't conceive of a development business where the maintainers rarely saw, let alone talked to, the developers. This last one happens when the offices are geographically isolated.

separated in space &/or time .... I once had somebody from the branch office that served AT&T longlines in NJ come looking for me because of some 10-15 year old kernel stuff.

in earlier times ... i liked implementing things in zero instructions (aka nominally purely as side effects of other things happening in carefully ordered sequence). years later ... somebody would make some random change and mess everything up ... and not have the slightest clue as to what was happening.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM says AMD dead in 5yrs ... -- Microsoft Monopoly vs. IBM
 monopoly
Newsgroups: alt.folklore.computers
Date: Tue, 04 Mar 2003 00:30:45 GMT

bdc@world.std.com (Brian 'Jarai' Chase) writes:
There's Cafe 50s on Santa Monica Blvd on the west side of Los Angeles. I'd expect there are a few classic soda fountain / diners to be found in nearly every big American city, and certainly in quite a few of the smaller ones.

dc pbs station had program on silver spring area diners from the 50s last week. one was the tastee diner in bethesda .... they showed it having its last serving before they broke ground for discovery channel hdqtrs ... and having to move it to its current location. tastee is still there ... re-opening not too long ago after being shut because of fire. i haven't been to tastee since it re-opened.
http://www.dinercity.com/mdDiner/mdDiners.html

tastee doesn't have soda fountain ... but there are the silver diners in the DC area ... which include juke box interfaces at most of the tables ... and you can punch in some amount of 50s, 60s, & 70s music. I've recently been to both the one on rockville pike and the one in tyson's.
http://www.silverdiner.com/

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

The PDP-1 - games machine?

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: The PDP-1 - games machine?
Newsgroups: alt.folklore.computers
Date: Tue, 04 Mar 2003 13:50:35 GMT

"Ross Simpson" <yeah_whatever> writes:
Careful, the creator of LISP maybe reading this. Naturally, I haven't used LISP, so cannot base an opinion on it.

some amount of drift:
http://www.mcjones.org/System_R/SQL_Reunion_95/sqlr95-Vera.html
https://www.garlic.com/~lynn/2002g.html#60

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Tue, 04 Mar 2003 14:06:06 GMT

eugene@cse.ucsc.edu (Eugene Miya) writes:
Lyons authored a fine annotation. I had a copy. However, I would not say that the best code always came from academia. Many fine OS features came from industry from both indutrial concerns (IBM and Tandem) and never saw the light of day, or from places like BBN. VMS, for instance, had features which created communities of use, whether Unix fans liked them or not. Berkeley style networking would look a lot like DECnet to a pure IBMer familiar with SNA or numerous other architectures behind the networking curve. Student code, is in many cases, far from perfect.

i think many people would take exception with SNA being referred to as networking ... it was communication not networking (aka large numbers, 30k, 60k, etc of terminals connected to mainframe). The first semblence of networking related to SNA was APPN ... and the SNA crowd non-concurred with APPN announcement. Eventually the APPN announcement letter went out but it was rewritten to eliminate any statement of connection between APPN and SNA.

when anne got con'ed into going to POK to be responsible for loosely-coupled architecture ... she originated peer-coupled shared data and got a lot of push back from the SNA crowd because of the "peer" part. there have been some references that large part of SNA was driven by pu4/pu5 interface for dealing with PCM (plug compatible manufactur) market (and cancelation of FS which had some possible objective for dealing with FS). And I've gotten some amount of blame for help originating PCM:
https://www.garlic.com/~lynn/submain.html#360pcm

and then later in HSDT to emulate SNA but using underlying peer network infrastructure:
https://www.garlic.com/~lynn/2003c.html#28 difference between itanium and alpha

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: alt.folklore.computers
Date: Wed, 05 Mar 2003 05:20:47 GMT

eugene@cse.ucsc.edu (Eugene Miya) writes:
The networking code in BSD was largely lifted from Rob's code at BBN. I tend to doubt that 1,000 implementations of TCP/IP were written. Copied: more than that. I am not certain by what metric you are claiming that it's best. A student at MIT has certainly implemented TCP in a much smaller space than previously thought possible (under 1KB), but that was only recently.

note that in the mid to late '80s it seemed like almost every unix was lifting their tcp/ip code from bsd ... i believe that the prevalent implementation on pc platforms was the mit tcp/ip code. various historical references:
http://www.ece.ubc.ca/home/staff/faculty/gillies/etc/www/9802net.html
http://www.apocalypse.org/pub/u/romkey/from.html
http://madhaus.utcs.utoronto.ca/local/internaut/pc-ip.html
http://www.ka9q.net/code/ka9qnos/

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Antiquity of Byte-Word addressing?

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Antiquity of Byte-Word addressing?
Newsgroups: comp.arch,alt.folklore.computers
Date: Wed, 05 Mar 2003 05:27:27 GMT

"Rupert Pigott" writes:
The more I learn about IBM the more I wonder if there is anything they haven't done...

there is the story about executive that had project in LA ... i think done in association with the (ibm) la science center and one of the university medical schools for blood dialysis machine .... and was involved in deciding whether to put a heater for the blood while it was out of the body and decided not to. some 25(?) years later when he was hooked up to one of his machines ... he was reported to have made some comment that cold blood was painful and not having a heater was possibly one of the worst decisions he had made.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

PDP10 and RISC

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: PDP10 and RISC
Newsgroups: alt.folklore.computers
Date: Thu, 06 Mar 2003 13:35:57 GMT

Pete Fenelon writes:
We're breeding a generation of applications programmers- and very few systems/tools guys to actually provide the bits. (Which, as a systems/tools sort of guy, reassures me, because it means those sort of skills will be increasingly at a premium as a generation raised on handholding environments teaches another generation.... ;))

the other way of looking at it ... is that was critical path to deliverables. in the '90s companies were paying obscene amounts of money to people to get applications into the market as quickly as possible ... and frequently the quickest way to demonstrate a deliverable was with scripting and lots of re-use. it hardly matter if it couldn't scale ... somebody would just pay obscene amounts of money for hardware, and if it got hacked ... well, getting hacked was intrinsic characteristic of computers ... nobody could do anything about it.

if one was so inclined, one might make an analogy to later periods in egypt ... it was easier to get building materials by scavenging off the pyramids.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Reviving Multics

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Reviving Multics
Newsgroups: alt.folklore.computers,alt.os.multics
Date: Thu, 06 Mar 2003 14:03:04 GMT

Tom Van Vleck writes:
"Having rings" may be a good start but one would have to look at the ring implementation in detail to see if the Intel, etc rings do the kind of things Multics likes.

Multics on the 645 had 64 rings, but had to simulate a lot of behavior that the hardware didn't support. Multics on the 6180 architecture had 8 rings, 0-7. On both architectures, the ring brackets protected segments. What would they protect on a flat machine? Needs study.

one might claim that access registers introduced as sort of follow-on to dual address space (which showed up in 3033). access registers provide it at address space boundary (rather than segment boundary) along with the capability of instructions to access different address spaces.

os/360 had started with lots of supervisor services that actually was program library loaded as part of application code and accessible by direct subroutine call in the same address apace ... and a pointer passing call paradigm. dual-address space was sort of addressing the paradigm that all kernel, supervisor services, library code, and application code all residing in the same 16mbyte (24bit) address space ... and all tied together with pointer-passing paradigm. dual-address alleviated some of the constraints by moving some supervisor services type stuff into another address space ... while still preserving pointer-passing paradigm (with some special tweaking that allowed supervisor services to use pointers that reached into different address space).

PC instrucation and access registers then sort of generalized it with some rules about what happened when calling various levels of supervisor services in other address spaces. Effectively, paradigm set up with different privileges enforced by hardware ... analogous to rings ... but across address spaces rather than across segments/rings. The justification for access registers was that 1) not significantly impact the pointer-passing paradigm that was intrinsic across the infrastructure, 2) allowed the efficiency of direct subroutine linkage with PC instruction and 3) support various kinds of privilege enforcement w/o the overhead of supervisor call where privileges were enforced by inline kernel instructions.

extracts from table of contents giving sort of evolution:
https://www.garlic.com/~lynn/98.html#36 What is MVS/ESA?

move detailed extracts of detailed description:
https://www.garlic.com/~lynn/2002d.html#51 Hardest Mistake in Comp Arch to Fix

small bit from above extract ... in some sense hardware preset for seven different (31-bit) address spaces (370-xa introduced 31bit addressing with 3081, esa/370 introduced access registers with 3090)
3.8.1 Changing to Different Address Spaces

A program can cause different address spaces to be addressable by using the semiprivileged SET ADDRESS SPACE CONTROL instruction to change the translation mode to the primary-space mode, secondary-space mode, access-register mode, or home-space mode. However, SET ADDRESS SPACE CONTROL can set the home-space mode only in the supervisor state. The program can cause still other address spaces to be addressable by using other semiprivileged instructions to change the segment-table designations in control registers 1 and 7 and by using unprivileged instructions to change the contents of the access registers. Only the privileged LOAD CONTROL instruction is available for changing the home segment-table designation in control register 13.

misc other dual address space and access references:
https://www.garlic.com/~lynn/2000c.html#84 Is a VAX a mainframe?
https://www.garlic.com/~lynn/2000d.html#28 RS/6000 vs. System/390 architecture?
https://www.garlic.com/~lynn/2000e.html#58 Why not an IBM zSeries workstation?
https://www.garlic.com/~lynn/2001d.html#28 Very CISC Instuctions (Was: why the machine word size ...)
https://www.garlic.com/~lynn/2001h.html#73 Most complex instructions
https://www.garlic.com/~lynn/2001i.html#13 GETMAIN R/RU (was: An IEABRC Adventure)
https://www.garlic.com/~lynn/2001k.html#16 Minimalist design (was Re: Parity - why even or odd)
https://www.garlic.com/~lynn/2002g.html#17 Black magic in POWER5
https://www.garlic.com/~lynn/2002g.html#18 Black magic in POWER5
https://www.garlic.com/~lynn/2002l.html#51 Handling variable page sizes?
https://www.garlic.com/~lynn/2002l.html#57 Handling variable page sizes?
https://www.garlic.com/~lynn/2002n.html#58 IBM S/370-168, 195, and 3033
https://www.garlic.com/~lynn/2002n.html#74 Everything you wanted to know about z900 from IBM
https://www.garlic.com/~lynn/2002q.html#1 Linux paging
https://www.garlic.com/~lynn/2003c.html#13 Unused address bits

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Filesystems

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Filesystems
Newsgroups: comp.sys.cbm,comp.sys.apple2,rec.games.video.classic,comp.os.cpm,alt.folklore.computers
Date: Thu, 06 Mar 2003 14:55:25 GMT

jeffj@panix.com (Jeff Jonas) writes:
I consulted to IBM for an AIX project. IBM fostered a horrible "not invented here" mentality. IBM had hired Convergent to make 2 AIX ports: for the PS/2 and the 3090 mainframe. They were tightly coupled into clusters and the TCF (totally confusing facility). No matter how much I wanted to dislike AIX, I saw many merits and unique features. I suspect that IBM killed that project (even though they owned it and PAID for it) since they wanted to re-invent it all with "native" code. Surprisingly, AIX for the Risc-6000 succeeded since it was written by the Austin Texas facility and was apparently exempt from the NY politics.

there were several ... totally different unix projects.

there was the tss/370 which modified the tss kernel so that at&t unix layer sat on top .... I don't know if that ever saw deployment other than inside of at&t. There was also a lot of effort to do a similar at&t port on top of VM ... but I don't know if that ever saw any customer installations.

there was the (at&t) pc/ix port by interactive for the pc. when the opd romp displaywriter project was killed ... the austin group quickly retargeted for unix workstation and got interactive to do a port. however the claimed that it would be faster if interactive ported to an abstract machine layer (rather than interactive having to learn the details of the romp hardware) ... and so a lot of the pl.8 displaywriter people were put to work writing the vrm (in pl.8). This was shipped (combined vrm & interactive) as aix for pc/rt (reworked displaywriter).

palo alto had been working on bsd port for the 370. with the advent of the pc/rt ... they retargeted their work to the pc/rt ... which eventually resulted in something called AOS (bsd for the pc/rt). They also somewhat disproved the original austin assertion having done a bsd port to the native pc/rt in possibly 1/10th the time/effort it took interactive to port to the VRM abstraction (and possibly 1/100th the time/effort for the combined interactive port plus VRM development).

AIX for the RS/6000 had AIX for the PC/RT reworked .... eliminating the pl.8 code and calling it AIX V3.

Palo Alto ... dating back to the early '80s had been doing work with both Berkeley and BSD as well as UCLA and Locus. In the early '80s Palo Alto had Locus running on S/1 and some 68000 machines in distributed Locus environment. This was ported to 370 and PS/2s and was the officially shipped AIX/370 and AIX/ps2. This palo alto/locus was totally different "aix" than the austin "aix".

Austin for AIX V3 had developed the journal file system where the filesystem was slightly re-organized so that all of the metadata was in area of virtual memory that was sort of designated database memory. This was sort of from one of the early 801 design activities where they wanted to show hardware assit for transactional database systems. Basically hardware assist to track transaction lines (about the size of cache lines) got modified ... and then at commit ... the commit code could run around and find all the modified lines/data and log it. Then you start the writes ... and if there was a failure ... restart could use the log to consistently roll-forward the database.

Early in the rs/6000/aix time-frame, Palo Alto had a project to port AIXv3 to different platforms. One of the inhibitors was the JFS dependency on the 801 database memory hardware. Palo Alto undertook to go thru the unix filesystem code and insert "log" calls whenever metadata was being modified (these "log" calls weren't necessary in 801 since all instances of data modification was magically identified by the hardware). It did have the downside with their relationship with austin (which was already strained .... in part because of showing them up with the bsd/aos port on the pc/rt ... as well as the locus activity shipping as aix/370/ps2) ... since the "log call" version was measurably faster than the original version using the 801 hardware. turns out in the hardware version, commit had to scan the lines in the database region for modified lines ... if the total size of metadata was significantly larger than the nominal lines modified per commit, the commit scanning took much longer than the overhead of the direct log calls.

For OSF there was an attempt at merge of many of these different threads ... in part for DCE ... bringing together mit, cmu, berkeley, ucla with various apollo distributed support, cms andrew distributed, the locus distributed stuff, austin aix distributed filesystem, etc.

aixv2 .... interactive at&t on top of vrm for pc/rt aos .... bsd for pc/rt aixv3 .... aixv2 with vrm eliminated for rs/6000 aix/370 .... ucla locus aix/ps2 .... ucla locus

there were misc. other bsd & at&t ports along the way for vm/370 and tss/370.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Another light on the map going out

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Another light on the map going out..
Newsgroups: bit.listserv.ibm-main,alt.folklore.computers
Date: Thu, 06 Mar 2003 15:53:33 GMT

thinkers_corner@HOTMAIL.COM (Thinkers Corner) writes:
10 years ago, IBM Management had told it is upto you to manage your career...

heck in the very early '70s i was told that i wouldn't have a career unless i joined FS (future systems) ... it was apparently unheard of to not join if asked. random fs postings:
https://www.garlic.com/~lynn/submain.html#futuresys

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Another light on the map going out

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Another light on the map going out..
Newsgroups: bit.listserv.ibm-main,alt.folklore.computers
Date: Thu, 06 Mar 2003 16:07:09 GMT

R.Skorupka@BREMULTIBANK.COM.PL (R.S.) writes:
OK, visit Unisys web pages. They have mainframes ...on Windows. This is big PC with Intel Xeon inside (couple of them). BTW: AFAIK they DID mainframes in the past, didn't they ? Unix based machines for sure, because i saw several.

in fact, I believe unisys started off this (mainframe) product line by re-logo'ing sequent boxes. during the '90s, sequent people several times claimed that all(?) of the (important?) parallelization work in NT was done on sequent boxes by sequent people (in any case the sequent line ran both NT and dynix). of course, sequent is now a unit of ibm.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Another light on the map going out

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Another light on the map going out..
Newsgroups: bit.listserv.ibm-main,alt.folklore.computers
Date: Thu, 06 Mar 2003 22:37:34 GMT

rhawkins@SINGNET.COM.SG (Ron and Jenny Hawkins) writes:
A HP Superdome, or SUN E15K I can accept as a mainframe class box (after all SUN got the technology from Cray). Not a V class.

Unisys, Burroughs, Honeywell and HP have all built mainframe boxes in the past, but I think I could consolidate some of those old systems onto my desktop and get faster and more reliable throughput.

some more drift ... unisys relogoed sequent.

both sequent and data general had done scalable multiprocessor ... 64-port SCI ... with 4 intel processors on each port for 256 processors.

convex exampler did 64-port SCI with 2-way HP risc on each port for 128 processors. HP bought convex and marketed under HP name.

the guy that did live oak (four rios .9 chips) was brought in to do superdome (long ago and far away he had gone with steve chen to chen supercomputers). we had some number of discussions of superdome vis-a-vis the convex exemplar.

when we started ha/cmp .... and fiber channel scale-up ... minor ref:
https://www.garlic.com/~lynn/95.html#13

we reported to executive that had come from motorola. be became head of somerset when that was formed (ibm, motorola, apple, etc) and then went on to be president of MIPS (sgi's machines).

other tidbits ... before IBM picked up Sequent ... chen was sequent's CTO.

there has sporadically been this joke that there are actually only 200 people in the business.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

POWER hashes vs tree

Refed: **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: POWER hashes vs tree
Newsgroups: comp.arch
Date: Thu, 06 Mar 2003 22:10:50 GMT

nmm1@cus.cam.ac.uk (Nick Maclaren) writes:
If I recall, ICL had been there already ....

so had future system ...
https://www.garlic.com/~lynn/submain.html#futuresys

the folklore being that after future system was canceled, some number of people moved to rochester to do s/38.

and before that tss/360 (on the 360/67)

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: alt.folklore.computers
Date: Thu, 06 Mar 2003 23:58:39 GMT

Peter Flass writes:
Right. It was the research universities that pushed networking to assist faculty collaboration. To take an example from another protocol, look at Bitnet. It was initially a Yale-Columbia[?] link that grew exponentially. If nothing else, cost kept it to the upper echelon at first, since each school had to pay for their own link.

vnet/rcsc started at the cambridge science center .... originally out of a project that connected the 1130 to the 360/67 (cpremote). one of the first production networking long haul links was between csc (cambridge) and endicott. This supported a joint (software) development project that implemented virtual 370 machines on cp/67 (running on 360/67) before 370 hardware was built.

this was all internal ... and slightly after the ncp/imp stuff on arpa. however the one thing that the vnet/rcsc got right was the layer between lower levels and the upper level .... effectively having the equivalent of gateway functionality from the start. one of the things that allowed the internal network to expand at much greater rate than the arapanet was the functionality ... and by the time of the transition of the arpanet to internet (1/1/83) and the start of bitnet .... the internal network was about four times larger than arpanet (and much larger than bitnet).

vnet/rscs for customer product got announced as part of a joint JES2/NJE/RSCS networking product in 1976. rscs then shipped with the JES2/NJE drivers .... which were much lower performance and functionality than the "native" (aka internal) rscs drivers. also jes2/nje got some of the layering wrong and intermixed things in the header fields that shouldn't be there. as a result, native jes2/nje systems at different version levels of the product didn't interoperate very well ... and in fact version/header incompatibilities could result in not only jes2 failures ... but the failure of the whole MVS system. On the internal network .... JES2s were typically relegated to boundary nodes ... with intermediate RSCS nodes with custom drivers on links to specific versions of JES2. It was the responsbility of the RSCS intermediate nodes to perform canonical rewrites of JES2/NJE headers (to prevent some JES2 somewhere in the world at one version from crashing some other JES2/MVS somewhere else in the world operating at a different version level).

Some of bitnet costs .... and all of the european "bitnet" (aka EARN) costs were subsidized by ibm. While a lot of the links were 9600 that was something of a cost issue (not being underwritten like a lot of the arpanet 56kbit links). There were numerous internal "campus environment" links that operated over ctca adapters.

When we started our skunkworks in the early '80s and the HSDT activities we got hyperchannel support .... which could more than match any speeds that the IMPs were capable of ... and were supporting both terrestrial and satellite T1 long haul links (and 50mbit local rates). I did the RFC1044 support providing RSCS as well as TCP/IP support over hyperchannel and NSC routers.

We were driving clear-channel long haul (terrestrial & satellite) T1 links with both RSCS and TCP/IP .... when the nsfnet backbone was installing 440kbit links (driven thru PC/RT router boxes). At the time of the NSFNET bid submissions ... a NSFNET audit of the backbone we were operating was deemed at least five years ahead of all bid submissions (not just the winning bid, to build something new from scratch).

misc bitnet/earn posts:
https://www.garlic.com/~lynn/subnetwork.html#bitnet

misc hsdt/hyperchannel posts:
https://www.garlic.com/~lynn/subnetwork.html#hsdt

misc. specific internet relates posts:
https://www.garlic.com/~lynn/internet.htm

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Bitnet again was: unix

Refed: **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Bitnet again was: unix
Newsgroups: alt.folklore.computers
Date: Fri, 07 Mar 2003 14:01:50 GMT

eugene@cse.ucsc.edu (Eugene Miya) writes:
With Bitnet you got 9.6 Kbps between mainframes. 11 years earlier one's institution paid from it's ARPA grants (a stiff requirement), $100K/year for 56Kbps. And that was NOT email.

I have some recollection in the late '70s somebody making reference that arpanet also needed 56kbits just to have 9600 available because of poor scaling and the inter-IMP chatter (he may have been slightly exaggerating?). recent thread:
https://www.garlic.com/~lynn/2003c.html#42 difference between itanium and alpha
https://www.garlic.com/~lynn/2003c.html#47 difference between itanium and alpha

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

Another light on the map going out

Refed: **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: Another light on the map going out..
Newsgroups: bit.listserv.ibm-main
Date: Fri, 07 Mar 2003 14:14:32 GMT

rhawkins@SINGNET.COM.SG (Ron and Jenny Hawkins) writes:
I still have friends that swear VMS is a mainframe operating system, just as good as S/370. Probably true, but who is running S/370? All you gals and guys seem to be running OS390 or zOS.

slightly related thread drift with vax & 4341
https://www.garlic.com/~lynn/2001m.html#15 departmental servers
https://www.garlic.com/~lynn/2002i.html#30 CDC6600 - just how powerful a machine was it?
https://www.garlic.com/~lynn/2002f.html#0 Computers in Science Fiction
https://www.garlic.com/~lynn/2002k.html#1 misc. old benchmarks (4331 & 11/750)
https://www.garlic.com/~lynn/2002k.html#3 misc. old benchmarks (4331 & 11/750)
https://www.garlic.com/~lynn/2003d.html#33 Why only 24 bits on S/360?
https://www.garlic.com/~lynn/2003d.html#35 Why only 24 bits on S/360?

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

ARPAnet again: Bitnet again was: unix

Refed: **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: ARPAnet again: Bitnet again was: unix
Newsgroups: alt.folklore.computers
Date: Fri, 07 Mar 2003 21:17:09 GMT

eugene@cse.ucsc.edu (Eugene Miya) writes:
I'd have to snag Taylor again. That was a decision made back in 1968. It might have had something to do with the tariffs. I doubt that chatter was a big deal.

somebody did reply in the previous referenced thread attesting to specific known instances where inter-IMP chatter could tie up the bandwidth. Mentioned was some negative feedback pathelogical cases where up/down state transition contributing to heavy chatter activity .... and that heavy chatter could exacerbate the perception of up/down state ... which would in turn, further exacerbate inter-IMP chatter.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

ARPAnet again: Bitnet again was: unix

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: ARPAnet again: Bitnet again was: unix
Newsgroups: alt.folklore.computers
Date: Fri, 07 Mar 2003 21:37:23 GMT

... and that as the nodes & links increased ... the probability significantly increased that something, someplace might experience an up/down state change (scaling issue independent of address space limitation).

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

IBM was: VAX again: unix

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: IBM was: VAX again: unix
Newsgroups: alt.folklore.computers
Date: Sat, 08 Mar 2003 04:19:40 GMT

Peter Flass writes:
Probably not too many S/1's left. Too bad, I thought the machine had a lot of potential, but I believe IBM was afraid to beef it up enough to compete with the PDP-11 for fear of cutting into the market for some of their other higher-margin producta.

I think most 4341 users have probably upgraded to larger mainframes. The same-size user now would probably be using NT or a unix system. IBM is hoping to catch the unix users as they grow with Linux/390.

many of 4341s went to customers in lots of multiple hundreds .... effectively departmental servers/time-sharing machines .... I believe that lots of these went to large unix & pc servers along with workstations & pc desktops ... rather than to the 4381 (4341 follow-on). A lot of the 4341s went to customers that were starting in the late '70s to migrate workload away from the monolithic glasshouse environment (STL was taking over a conference room on every floor in every tower to be a local 4341 room).

S/1 were 64k address machines and in the mid-80s (had been around for 15 years) going thru some number of hoops to come up with more than 512k byte real memory. Part of the S/1 issue was similar to 4341 ... the predominate market was quickly going to large PCs and workstation products in the mid-80s. 4955 mdl e had max. storage size of 256k bytes and 4955 mdl f had max. storage size of 512k bytes.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: alt.folklore.computers
Date: Sat, 08 Mar 2003 15:08:11 GMT

Brian Inglis writes:
Personally, I did not see it as a big deal: in those days of hex dumps, even on VM, reading thru a few K of code was not uncommon, particularly with third party binary products.

note too long after i gave presentation on dumprx (a binary product) at share ... i got a number of requests regarding some of the implementation details ... and shortly there after a number of 3rd party binary products showed up on the scene with feature/functions I had described in share presentation.

i was directly distributing dumprx to something like 100 internal locations and I was told all field support/PSRs used it (having acquired it in one way or another).

Shortly after REX(X) appeared .... I undertook to do dumprx as a demonstration that REX was not just another scripting language. The existing IBM binary product, IPCS was written in assembler, something over 20k instructions and had a support group of something like 8-10 people. The dumprx demonstration was that I could in 3 months elapsed time (less than 6 person weeks) develop something that 1) ran ten times faster (aka that interpreted REX could run ten times faster than assembler), 2) had ten times as much function as IPCS, and 3) the implementation code had to ship with the product.
https://www.garlic.com/~lynn/submain.html#dumprx

once the basic infrastructure was in place ... i started a library of scripts that would analyze memory for most of the typical fault signatures (this was included in the original time constrained objective).

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

Refed: **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Sat, 08 Mar 2003 15:29:41 GMT

jmfbahciv writes:
Nope. Another reason IBM would buy a VAX is to test their software if it had to talk to a VAX. Even computer manufactureres bought computers based on the job it had to perform. For years, DEC used a Burroughs to do their payroll.

my brother (at the time regional marketing for apple) talked about using his appleII to dial into the hdqtrs s/38 to check on ships, schedules, and other things. he claimed that he had some advantage over most of the other marketing types having come up from a technical background ... and figured out how to get into and use the hdqtrs data processing machine (somewhere along the line i believe they upgraded from s/38 to as/400).

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Sat, 08 Mar 2003 15:53:20 GMT

jmfbahciv writes:
Nah, I wouldn't assume that. Most IBM interfaces had to suck (sorry, Lynn). I don't know why but it always seemed to take awful code written for a very long time to do anything IBMish and that included talking to their systems. I have no idea if the reasons were because IBM sucked or the corporate cultures of the two computing styles clashed or what.

the "official" communication interfaces were SNA ... there has been some folklore that the complexity of those interfaces are result of project i was on as an undergraduate that has been blamed for originating the PCM (plug compatible manufactur) market.

Originally, in theory, FS was suppose to address that ... with extremely advanced, sophisticated integration. When FS got canceled, the individual product lines were left on their own. The folklore sort of indiciates that complexity was then substituted for sophisticated.

there was joke (from a number of places) that if an internal product attempted to jmplement sna support according to the official internal sna specification it would never work (boca/s1 was one group that complained bitterly about it). the only way to make something really work was reverse engineering & regression testing with real pu4/pu5 operation.

For a period, my wife was chief architect for amadeus (airline res system for europe, and a couple us lines). she settled on x.25. the internal sna crowd created such an uproar that they got my wife removed. it didn't do much good ... amadeus went x.25 anyway.

earlier she had been conned into going to pok to be in charge of loosely coupled architecture .... where she originated/authored Peer-Coupled Shared Data architecture ... which was totally counter to non-peer SNA paradigm. there were frequent battles between how far glasshouse datacenter operation could extend peer-to-peer before they had to switch to non-peer sna. as fiber technologies appeared ... the processor group kept trying to push peer-to-peer into multiple kilometer range before they had to officially bow to sna non-peer.

there was one research project that used 8-tail trotter (3088 switch with eight channel-to-channel connections) to support an eight cec cluster. they initially developed a peer-to-peer cluster synchronization protocol that would settle in under a second. They then had their arm twisted to layer it on top of sna using a non-peer-to-peer paradigm .... which elongated the sub-second cluster synchronization settling to a couple of tens of seconds.

misc future system:
https://www.garlic.com/~lynn/submain.html#futuresys

misc pcm
https://www.garlic.com/~lynn/submain.html#360pcm

misc saa, sna, &/or 3 tier
https://www.garlic.com/~lynn/subnetwork.html#3tier

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

Refed: **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: alt.folklore.computers
Date: Sun, 09 Mar 2003 15:21:21 GMT

bdc@world.std.com (Brian 'Jarai' Chase) writes:
A lot of the companies producing commercial OSes of that era allowed customers access to the source, either on fiche or in machine readable source. A few examples being the IBM products you've listed above, VMS, TOPS-10, TOPS-20, Multics, and probably lots of others. My specific question is looking to find out about any other companies which allowed their customers to resell the OS, with modifications, as their own product.

AFAIK, no one else apart from AT&T did this, and from what I understand, the only reason AT&T did this was because of their unique situation, being regulated by the government.

operating systems didn't use to be licensed. in theory anybody could pick the operating system up and use it. that is what led to the pcm (plug compatible manufactur) cpu market. operating systems can be extremely expensive to develop ... as hardware technology evolved ... the cost of developing operating system could be significantly larger than the cost of developing a processor.

i helped develop a PCM controller while an undergraduate and have gotten blamed for originating the PCM controller market. however, PCM processors didn't show up until a number of years later ... although Amdahl made no secret of the fact that he was developing one.

A lot of change-over to restricted operating system licensing and then later to actual separate pricing for the operating system was a result of the appearance of the PCM processor market. Along the way, there was some amount of litigation ... and at least one settlement I believe resulted in what you describe ... alhtough with the appropriate fines and licensing fees.

A trivial, early one is CP/67 and the break-away group that formed NCSS and called their product (something else?) ... although most of the NCSS installations were internal to NCSS offering NCSS timesharing service. I sometimes somewhat facetiously pick on multics ... based on the number of actual customer installs ... making claims like that very small number of machines running NCSS (compared to the total cp/67 installs) possibly approached the total number of customer installations running multics. I know that at various times when I was personally creating production vm/370 distributions that I would directly mail distribution tapes to small number of internal sites that was larger than the total number of customer machines running multics.

In the who's bigger contests ... it wasn't fair to quote total intenal installs, or total external customer installs .... I would just quote the numbers that I personally supported against the total multics installs. It was somewhat friendly since both CP/67 and Multics was going on in 545 techn. sq .... but on different floors in the same bldg.
https://www.garlic.com/~lynn/subtopic.html#545tech

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: alt.folklore.computers
Date: Sun, 09 Mar 2003 16:18:33 GMT

Brian Inglis writes:
IBM mainframe systems all came with source on microfiche to study and modify: DOS (later VSE), OS/VS1 (SVS), and OS/VS2 (MVS) were basically binary distributions AFAIK, although maybe the machine readable source was available, perhaps for a charge, possibly large.

minor footnote #1

frequently microfiche was used to get a first cut at a possible superzap. there is this story about some gov. agency requesting microfiche that exactly matched their MVS (kernel) running code. After much investigation (costing a reported $6m), the answer was NO. Part of the problem is that the microfiche is cut very early in the intergration and test cycle (basically at about the time that documentation stuff goes off to the printer) ... and all sorts of stuff happens later.

minor footnote #2 ....

os/vs1 started out with effectively mft layed out in 16mbyte virtual address space. It was somewhat as if you were running mft in a 16mbyte real machine (with some low-level stuff to mask page faults and other stuff from the rest of the code).

os/vs2 started out with effectively mvt layed out in 16mbyte virtual address space. again think mvt running in a 16mbyte real machine with some low-level stuff masking page faults.

os/vs2 had a two stage plan ... start with SVS .... a single 16mbyte virtual address space .... and then do the additional development to later offer an implementation where each application got its own 16mbyte address space (aka MVS).

from
http://os390-mvs.hypermart.net/mvshist2.htm MVS history
OS/VS1 provided a single virtual storage address space system, while OS/VS2 allowed multiple virtual storage address spaces. However, the first release was restricted to a single virtual storage address space and became known as OS/ VS2 SVS. The following release, made available in July 1974, contained multiple virtual storage, address space support and was named OS/VS2 MVS Release 2. Both OS/ VS1 and OS/VS2 SVS supported a total of 16MB of virtual storage. Because the OS/VS2 MVS release supported multiple virtual storage address spaces, each of which provided 16MB, most people assumed it would be years before additional storage would be required.

... snip ...

note that to preserve the OS/360 pointer passing paradigm ... the kernel code resided in the address space. In MVS, half of each address space (8 mbytes) was for kernel and half (8 mbytes) was available for application .... modulo something that become to be called the common area ... basically something that came out of the application 8mbyte, which could be two megabytes or more ... leaving 6mbytes or less address space for applications. The kernel and common area was the same in all address spaces. This design point would result in a signficant hardship for vm/370.

The 168-3 increased the processor cache to 64kbytes. As part of the direct cache mapping they decided to use the 8mbyte address bit. However, it was only active when running in 4k virtual page mode. When the machine ran in 2k virtual page mode, the additional cache lines were disabled. Anytime the machine switched between 4k virtual page mode and 2k virtual page mode, the complete cache was flushed. VS/1 ran with 2k virtual page mode, and when VM was running VS/1 in a virtual machine, it ran it with 2k "shadow" tables. A vm/370 customer that had a lot of production vs/1 or dos/vs and upgraded from a 168-1 to a 168-3, not only didn't see any performance improvement, but actually would see a significant performance decrease.

The "Common area" was for subsystem implementations that were moved into their own "private" address space but still needed to preserve pointer-passing paradigm (so data could be directly addressed). This was somewhat address later in the dual-address space support in the 3033 .... allowing semi-privileged code (subsystem in its own private address space) to access data using pointer from (different) application address space. Part of the dual-address space was to alleviate the common area which was starting to limit some installations to 4mbytes for applications.

slight trivia ... person that worked on dual-address space as well as on doing some work on first 32-bit 801 (blue iliad) chip, then went on to join HP and work on HP's risc chip (there is this folk tale that he had given notice and was going to start with HP on january first .... but he spent nearly round the clock the last two weeks of dec. working on blue iliad).

past refs:
https://www.garlic.com/~lynn/2000c.html#84 Is a VAX a mainframe?
https://www.garlic.com/~lynn/2002g.html#18 Black magic in POWER5

slightly related joke about only being 200 people in the industry:
https://www.garlic.com/~lynn/2002.html#16 index searching
https://www.garlic.com/~lynn/2002l.html#37 Computer Architectures
https://www.garlic.com/~lynn/2002o.html#9 PLX
https://www.garlic.com/~lynn/2003d.html#57 Another light on the map going out

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

DOS trivia question

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: DOS trivia question
Newsgroups: alt.folklore.computers
Date: Sun, 09 Mar 2003 14:56:21 GMT

i've got about 50 5.25in floppies that i no longer have floppy reader for. at least four or five of them are labeled with various versions of dos1.8.

dos1.8 was an internal only version .... with large number of bells and whistles. one of the things was generalized rewrite of the diskette handling interface with configurable interface that had load of options. I used it with a PC configuration that had two normal 40-track dual-side diskette drives and two external TEAC 80-track dual-side diskette drives. I could format all diskettes at 10 sectors per track.

dos1.8 was done by a specific person .... but I can't remember his name (I could probably get it if I was able to read the floppies). Does anybody remember the name of the person responsible for dos1.8?

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

SSL/TLS DHE suites and short exponents

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: SSL/TLS DHE suites and short exponents
Newsgroups: sci.crypt
Date: Sun, 09 Mar 2003 16:56:08 GMT

daw@mozart.cs.berkeley.edu (David Wagner) writes:
You trust Verisign to check whether q is prime? Remember that each cert-holder could be using a different set of parameters.

I doubt any commercial CA is set up to do this checking, and I'm not sure I'd trust them to do it properly anyway.

for the most part TTP CAs are certifying that

1) you own the private key that corresponds to the presented public key ... possibly by using the public key to verify some digital signature ... modulo a reply attack

2) that the information you assert which is to be bound in the certificate is valid

in the case of a ssl domain name server certificate ....

a) i assert that i'm the owner of the private key that corresponds to the presented public key

b) i assert that i own the presented domain name.

so the TTP CA stuffs the presented public key and the presented assertion into a digital certificate and signs it (after doing a little verification).

checking the quality of a public key and/or a public/private key pair is a separate business process from the typical certification. There are gobs of things that might affect the quality of the key pair ... like have both the public and private key been publicly disclosed in some way. From a business perspective such things are also on par with possibly certifying the quality of web site .... as opposed to certifying the ownership of a domain name.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

cp/67 35th anniversary

Refed: **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: cp/67 35th anniversary
Newsgroups: alt.folklore.computers
Date: Mon, 10 Mar 2003 14:06:33 GMT

cross-posted from vm-esa.

1968 spring houston share was the first week in march. this year's 2003 spring share was in dallas, feb 23-28 (this probably isn't a long term url):
http://www.share.org/dallas

i was undergraduate at university when dick bayles, hurit nanavati, and john harmon came out from cambridge science center the last week of january to install cp/67 .... making the university the 3rd cp/67 installations after cambridge itself and lincoln labs. then at the spring houston 1968 share meeting i got invited to come along and be part of the "official" cp/67 product announcement (35 years ago).

june of 1968 there was an ibm customer education class for cp/67 held in beverly hills. i think that friday before .... bayles, et al had left and formed ncss (first cp/67 time-sharing bureau). dick was supposed to be one of the primary people teaching the class, but because he had just left the company he was not allowed to. I was one of the people that got roped in at the last minute to try and teach the material (while still an undergraduate).

at the university i got to do a lot of fiddling with both os/360 and cp/67 (even some tss/360). one of the projects was when I discovered I couldn't do what i wanted to with the 2702 .... was project for the first telecommunication controller. we took interdata/3 and reverse engineered the ibm channel .... and built the first pcm controller .... in the process getting blamed for originating the pcm controller business. the objective was to be able to do both dynamic baud rate determination as well as dynamic terminal type determination (across 2741, 1052, & ascii/tty terminals).

cp/67 & mft14 report i presented at fall Atlantic City share meeting
https://www.garlic.com/~lynn/94.html#18 CP/67 & OS MFT14
https://www.garlic.com/~lynn/94.html#20 CP/67 & OS MFT14

minor recent pcm related threads:
https://www.garlic.com/~lynn/2003d.html#44 Why only 24 bits on S/360?
https://www.garlic.com/~lynn/2003d.html#49 unix
https://www.garlic.com/~lynn/2003d.html#67 unix

some recent interdata related threads:
https://www.garlic.com/~lynn/2003c.html#15 difference between itanium and alpha
https://www.garlic.com/~lynn/2003c.html#28 difference between itanium and alpha
https://www.garlic.com/~lynn/2003c.html#70 COMTEN- IBM networking boxes
https://www.garlic.com/~lynn/2003c.html#76 COMTEN- IBM networking boxes

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

Refed: **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Mon, 10 Mar 2003 14:31:37 GMT

jmfbahciv writes:
I wouldn't believe that folklore. You seem the type that, once the bits are set, you don't redefine them.

according to folklore (as well as recorded anecdotes), a primary driver in FS was tight, sophisticated intergraiton between all of the components to address the emerging PCM (plug compatible manufactur) market. When FS was killed, the responsibility went back to the individual box organizations. The folkloare and the appearance of the (SNA) pu5/pu4 interface (vtam/3705ncp) is highly integrated, extremely complex, and tightly controlled (which may also account for some of the non-peer-to-peer characteristic as well).

the project i worked on as an undergraduate has been blamed for originating the pcm controller market.

specific fs recorded anecdote:
https://www.garlic.com/~lynn/2000f.html#16 [OT] FS - IBM Future System

another anecdote was that the amount of money/effort sunk into (canceled) FS would have bankrupted any other (computer) company.

misc. fs postings:
https://www.garlic.com/~lynn/submain.html#futuresys

misc. pcm postings
https://www.garlic.com/~lynn/submain.html#360pcm

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: comp.os.vms,alt.folklore.computers
Date: Mon, 10 Mar 2003 15:52:45 GMT

Lars Poulsen writes:
And in the case of IBM Federal Systems, the project group would build a system on the foundation most likely to work. In some cases, that might be a VAX/VMS system even though one might speculate that a Series/1 or a System/34 could possibly be tweaked to do the job.

As a 3rd-party subcontractor, my employer was often invited to bid our VAX-compatible subsystems into an IBM-FSD bid for a Federal Agency. I was very impressed that IBM allowed FSD to bid based on the customer's best interest rather than on what maximized IBM's participation percentage.

FSD was more akin to a large (federal gov) system integrator ... that happened to be owned by IBM. There use to be a joke that the two largest activities at FSD was:

1) gml programming (aka writing RFI/RFP responses and contracts) 2) managing subcontracts

when we were doing some of the HSDT (aka high speed networking) in the mid-80s somebody suggested that it would be helpful that we get FSD input on what we were doing (since some of them had been involved in things related to high speed). Now we had already done things like talk to nasa people in houston that were trying to figure out what to do about the space station ... which was going to have a 100mbit/sec downlink of continuous data (aka there was a continuous 100mbit/sec of data that didn't stop). Turns out it wasn't really a network problem ... it was what to do with the firehose of bits coming at you.

In any case, we thot, what the heck. We put together a presentation and they reserved a room. FSD had 30 people show up and they dragged the presentation and Q&A out for a week. And then we never heard anything more .... other than we got a bill from FSD for a 30 weeks worth of people time at fully loaded rate (they were under gov. billing/accounting rules ... and they needed to bill 30 people for a week).

so the 3rd largest activity was probably

3) billing and accounting (conform to gsa audits)

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

unix

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: unix
Newsgroups: alt.folklore.computers
Date: Mon, 10 Mar 2003 16:00:58 GMT

jmfbahciv writes:
DEC frowned on at least one biz who took TOPS20., renamed it to MYMONITOR. and tried sell it. They were stupid enough not to run a TECO macro over all sources and look for all occurences of TOPS-20 in all its misspelt forms. They also didn't corrent the misspelts error messages.

the person who originally wrote RSCS .... a number of years ago had an opportunity to work on some real time sysstems (he was doing some stuff for FSD at the time). There is a propular monitor that is used by many of the embedded real time systems. He commented that he was somewhat surprised to find the core real-time task scheduler was an instruction by instruction translation of his 360 assembler into C, preserving all of the original comments (including any misspellings).

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

reviving Multics

Refed: **, - **, - **, - **, - **, - **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: Re: reviving Multics
Newsgroups: alt.os.multics,alt.folklore.computers
Date: Mon, 10 Mar 2003 18:10:54 GMT

westinnospam@graphics.cornell.edu (Stephen H. Westin) writes:
From the point of view of instruction decoding, it's simple and regular. Quite unlike, say, a PDP-11 where different instructions can

two sorts of simple & regular ... while there were/are 2-byte, 4-byte and 6-byte instructions ... the first two bits of the instruction opcode indicated the instruction length. 2-byte tended to be register-to-resiter, 4-byte tended to be register-storage or immediate-storage, and 6-byte tended to be storage-to-storage.

storage addresses were 12bit offset with base register, some four byte instructions included index register in addition to base register.

by the time of 370, the 8bit op-codes were starting to be used up, so B2 was selected as an extension with the 2nd byte now the opcode. it was always a four byte instruction ... as indicated by the first two bits of the B2.

part of the need for a principles of operation (and its superset, the architecture red-book) was that a broad range of different technologies and plant-site locations around the world were implementing to the 360 standard.

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

'Boyd': A military Strategist's Emphasis on Speed

Refed: **, - **, - **

From: Anne & Lynn Wheeler <lynn@garlic.com>
Subject: 'Boyd': A military Strategist's Emphasis on Speed
Newsgroups: alt.folklore.computers
Date: Mon, 10 Mar 2003 22:31:25 GMT

one of my favorite topics ... and somewhat computer related since for a time he ran one of the largest datacenters in the world ... aka
https://www.garlic.com/~lynn/2003.html#51 Top Gun

book review:
http://www.nytimes.com/2003/03/09/books/review/009SPECTT.html?ex=1048329104&ei=1&en=6ac9bc0ff7348440

what is missing in some of this with respect to OODA-loops is not only speed (as in 'maneuver warfare' and business competititon) but also being able to look at the subject from every aspect/facet.

misc. refs:
https://www.garlic.com/~lynn/subboyd.html#boyd

--
Anne & Lynn Wheeler | https://www.garlic.com/~lynn/
Internet trivia 20th anv https://www.garlic.com/~lynn/rfcietff.htm

next, previous, index - home