PROJECTED REQUIREMENTS FOR PROCESSING ON IBM-1401 AND IBM-1410, SEPTEMBER 1964--DECEMBER 1965.

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78-01092A000100100006-3
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 12
• Document Creation Date: December 12, 2016
• Document Release Date: April 8, 2002
• Sequence Number: 6
• Publication Date: September 2, 1964
• Content Type: MF

Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
2 September 1964
MEMORANDUM FOR: DAD/OCS
Chief, Operations Division
Chief, Plans Staff
SUBJECT Projected Requirements for Processing on
11314-1401 and IBM-1410, September 1964--
December 1965.
1. Chart 1 (attached) lists monthly hour estimates of
IBM 1410 utilization per job for selected months through
December 1965. The estimates for present jobs are based
on quite reliable factors and thus, at least in total, the
chart should reflect a fairly accurate but conservative
prediction of future requirements. A modest amount,-1-45
hours per month, is estimated for hitherto unidentified
applications. Maintenance time is not included and must be
considered as another. 40-50 hours per month which is not
available for usage by. OCS.
2. The totals on the chart (hours based on current
1410 configuration of 40K, 600 LPM printer, and MOD 5 tape
units) show a rapid increase in utilization from 309 hours
in June 1964 to 540 hours in December 1964. A less, but
still significant rate of increase, continues throughout
1965; and this lesser rate of increase very possibly%is a
reflection of the conservative estimate for future and yet
unidentified requirements.
3. Monthly computer usage figures in a given shop
must be interpreted with emphasis on considerations of the
job mix ratio of fixed schedule versus ad hoc machine runs.
In the extreme case of a shop with all neatly dovetailed
fixed scheduled runs, presumably all production requirements
up to 4?30 hours a month (allowing 48 hours maintenance)
could be met in a three-shift (528 hours) operation. In a
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
STATINTL
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
shop with ad hoc top priority requests, it is probable that
occasionally several of these ad hoc requests will be made
simultaneously, and thus, customers will be unhappy. In
OCS many priority jobs require a large number of hours and
thus the probability of an unhappy customer is greatly
increased. OCS must set a compromise policy of a maximum
average projected utilization time for our run and priority
mix. Projected usage greater than this average would
indicate some action to be taken, such as lowering priorities
of selected jobs or increasing the hardware capability.
4. Past production performance indicates that a monthly
utilization of 300 hours with third shift availability provides
reasonably adequate answering of ad hoc priority requests.
The acceptability of the production performance rapidly
decreases with an increase in the average utilization and
at 375 hours.per month, OCS can expect many unhappy customers.
Above 400 hours per month OCS will have a degenerating
situation in which progranuner/analysts will be wasting con-
siderable.time trying to provide partial answers to ad hoc
requests.
5. Recently IBM reduced the overtime rental to 30%
of prime time, and since this 30% in reality applies only to
.the central units, a large monthly usage produces a very
attractive computing power per dollar ratio. However, OCS
,cannot let this attractive overtime pricing be a crippling
-deterrent to the policy of answering priority ad hoc requests.
Future OCS hardware plans should assume a planned maximum
of no more than 300-330 hours per month.
6. Using the totals of Chart 1 as a base, Chart 2
:gives estimates of the projected monthly utilization in
hours per selected months for 13 possible hardware configura-
tions. Also given is the monthly hardware rental based on
these estimates and on current rental factors.
7. One obvious recommendation which can' be drawn from
.this chart is to lease the acceleration feature immediately.
In all cases it would reduce monthly rental per production
;;output, but most important it would give extra processing
hours.
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
8. If the hours per month per configuration estimates
and the above suggested monthly use maximums are valid, OCS
must begin upgrading its 1410 complex immediately. Even the
immediate ordering of the acceleration feature as an addition
to the orders for the 40K additional memory and the oneilOO
LPM printer, will not avert an OCS crisis by April 1965.
Two possible general configurations exist to upgrade the
IPM--1410 for the 1965 workload and'a choice and an order
should be placed now.
-.,a. One IBM 7010, 80-100K, Mod 5 or 6 tape units
giving a good capability of doing the projected
OCS workload for approximately $30,000 per month.
Various gradiations of components exist within
this system giving varying capabilities of meeting
priority ad hoc requests (see Chart 2).'
.b. Two IBM 1410's, 60-80K, Mod 5 tape units
giving an excellent capability for doing the pro-
jected workload and for meeting priority ad hoc
requests at a cost of $43,000--45,000 per month.
9. Recommendation of IBM 7010 Several jobs exist
which demand top priority ad hoc processing. However, unless
OCS can ascertain that the priority of at least one of these
jobs alone justifies the second IBM 1.410 as backup, an IBM
7010 must be recommended for lease in March 1965. The large
price difference could not be justified nor rationalized in
any other way.
The hardware is 100% compatible with our present
configuration. Our present software (OPS and non-oPS) is
100% compatible and software which would be written in OCS
for the 7010 would be largely compatible with most third
generation machines.
10. Peripheral Support to the IBM 1410 and IBM 7090.
Peripheral Support to the 1410 and 7090, which
is presently done on the 1401, consists largely of printing.
Also most production jobs which had been programmed for the
1401 are printer I/O bound. Both types of production
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
frequently are run on the 1410 to answer hardware malfunction
or peak load conditions on the 1401. The 1401 is now at
peak capacity and is utilized nearly 400 hours* per month.
At various processing points during the running of some large
priority ad hoc 1410 jobs, print image tapes are produced
which in practice must wait to be printed on the 1410.after
normal run completion because the 1401 is saturated with
printing from the 7090. This produces a linear sequence
rather than parallel processing. An additional print
thruput channel must be acquired to provide output in reason-
able time.
11. Optional Solutions for Extra Printing CaRability.
a. IBM 1460--Replacement of 1401 giving 1100 lines
per minute instead of 600 at approximately same
price. All hardware and software would be compatible.
(1) PRO--Nearly double thruput per dollar.
(2) CON--Only one channel for thruput. In
practice would greatly help 7090 turn-around
time, but do little for 1410. At least one job
scheduled for future will use 120 character
print chain which fits only on 1401.
b. ?IBM 1460 as an addition to hardware complex.
(1) PRO--Additional 1100 lines per minute
extra thruput. 1401 still available for 120
character print chain.
(2) CON--Cost ($8,000) is high. As an addi-
tional unit a very low turnaround time priority
would be required to justify such a high cost.
Very little of the extra computing power avail-
able could be utilized.
*Utilization = Clocktime/0.70
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
c.' Second printer added to 1.410.
(1) PRO--Additional print capability for
approximately $2,300.
:7(2) CON--Software to work this capability into
OPSYS is poor. Very probable that the net
print speed would be reduced to 200-400 lines
per minute. Also most certainly some thruput
-speed would be sacrificed on the simultaneous
.=processing programs. All factors considered,
.,additional printing per dollar would probably
be low and operational management would be
difficult.
d. IBM 1440-~-As an addition, composed of CPU, console,
TU, and 1100 LPM printer.
::t (1) 'PRO--Cost ($3575) very high thruput per.
dollar. The additional thruput channel would
alleviate multiple peak load conditions. Very
good capability for data transmission support
through addition of Tele-unit at $250. Soft-
t, ware is available and good.
:(2) :CON--Little capacity for computing backup
:.without. special programming.
e. - Stromberg- Carlsen 4020.
-.Converts magnetic tape digital signals into
alphanumeric printing, curve plotting, and line
drawings on microfilm and/or photo paper at high
speeds. By adding a Xerox Copyflow, hard copy can
be produced from the standard microfilm rolls.
Total equipment cost is $214,000 which equates to
roughly $6-7,000 monthly rental. (For comparison
purposes.) For further specifications and possible
uses see Memo for Chief, Applications Division;
Subject: Equipment to Reduce Costs of Producing
Computer Output; by R. 'Do Hicks; dated 9 March 1964.
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
 Approved For Release 2002/05113 : CIA-RDP78-01092A000100100006-3
(1) PRO--Microfilm and plotting capability
-possibly could be useful. Single hard copy
processing is comparable, being cheaper than
the 1460 or 1401 during prime shift (more
expensive during second and third shifts).
(2) CON--Turnaround time is slow for hard
copy. Cost is expensive compared to 1440
printing. 122,000 lines of hard copy cost
$40.00 ($20.00 per hour prime time) on 1440 and
$49.00 on SC-4020; however, extra shift rental
of 3.440 is approximately $5.00 per hour.
N-copy listings must be processed n-times with
the resultant further increase in cost.
f. Hone ell-200_(an addition).
Essentially a copy of the IBM 1401/1460,
reputedly more flexible and with better performance.
Cost for a stripped 2K system with TU and a 900-1260
LPM printer is $3,040 per month. This expanded to
a 4K system (add $250) and with the editing feature
(add $90) costs $3380 and allows use of the SCOPE
software package which handles all typical peripheral
usage..
(1) PRO--High processing power per dollar
ratio. Software is good and 1401 software can
be run in compatibility if H-200 is expanded to
8K. The print speed range of 900-1260 permits
more lines per minute than the IBM 1403-3 when
the character set used is mostly digits or blanks.
Additional computing power is available in the
4K memory. Present customers generally are
,,.-well pleased with overall performance.
.(2) CON--M-H hardware and software in past has
generally been considered to be inferior to that
of IBM. The present claims to performance are
being verified by new customers, but some residual
reluctance is prudent.
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
CDC-8090 (an addition.
The CDC-8090 is essentially a cheaper, more compact,
more flexible version of the reliable CDC-160A. CDC
now has its own 1.00-LPM printer which has been per-
forming reliably for 15 months. Quality of print is
good but slightly inferior to that of IBM 1403.
The CDC 8090 is one of the most adaptable peripheral
I/O devices for a larger computer on the market.
A.standard software package will run four 10QO LPM
printers and a card/punch simultaneously on unit
record image data (print controls can be used) at
the same time at full speed. Rental price for 4K-
8090 with one printer is $3445.
(1) PRO--High processing power per dollar
ratio. Software generally is good and specifically,
software for handling unit-record images ' is
excellent. Additional 1000 LPM capability can
be added for $1600 per month (maximum 4) with no
change in software. Very reliable performance.
(2) CON--Slightly more expensive than H-200
($65.00 per month). Printer output slightly
inferior to IBM 1403. Capability to process
non unit-record image data would require extra
programming.
12. Recommendation of. CDC 8090. Even though the IBM 1440,
H-200 and the CDC 8090 are very close in performance and price,
the CDC 8090 is recommended as first choice for the following
reasons:
a. Reliability--comparable to 1440 and a past per-
formance record superior to Honeywell in hard and
software. .
b. Flexibility. More flexible than H-200 or IBM-1440
in unit record image capability. An ever increas-
ing proportion (now 70%) of OCS peripheral printing
is unit-record image.
Approved For Release 2002/05/13 : CIA-RDP78701092A000100100006-3
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
as an extra.)
i. A data transmission capability
can be. added
to
CDC 8090 for $125.
j. The IBM
1401 printing thruput
channel in OCS
can remain,
thus meeting the future requirement
for
c. Speed is near top of today's market.
d. Software (for unit-record image).is the best.
e. An additional print thruput channel is provided
giving operations a much increased scheduling flexi-
bility. Up to three extra thruput channels (ex-
cluding CPU) provide for ease of future expansion.
f. Tape units are compatible with IBM units.
-g. The 90-120 day delivery time answers OCS require-
-ments.
h. OCS already has approved a CDC 8090 for a field.
-.installation. An 8090 for COMMO is. arriving at
Headquarters in October for program checkout, and
then will be shipped to field. Tentative and unofficial
COMMO plans call for a proposal for a second CDC 8090
field installation after appropriate program checkout
at Headquarters. A CDC 8090 in OCS available one
to two hours per day for COMMO program checkout would
cancel the requirement for the Headquarters COMMO
installation. (Paper tape-reader and punch required
a 120 character print chain.
k. Psychological impact on IBM. Probably the
greatest advantage to be derived from the lease of
a CDC 8090 (or other non-IBM gear) to be placed next
to'present IBM equipment would be the competitive
jolt given to the IBM sales, systems, and engineer
Approved For Release 2002/05/13: CIA-RDP78-01092A000100100006-3
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
forces. Direct comparisons of performance would
be valuable to OCS in making decisions about the
very important future: direction. of OCS hardware grow tATINTL
Distribution:
.._. Orig. - -- File
1 - Each addressee
1 -> chrono
ILLEGIB
TATINTL Applications
'4328/(24 September 1964)
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
 Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
ACTUAL & Projected Usage of OCS IBM - 1410
(in hours per month)
Compiled by 0 25 AugustSIACTNTL
Assumes current configuration
May
June
July
Sept
Dec Ma
r J
une
Sept
Dec
64
64
64
64
64 65
6
5
65
65
3 1
1
1
3
40
60
46
50
52 54
5
6
58
60
1
3
4
6 6
6
- 6
6
3
20
15
10
1
1
1 1
1
2
10
2
2 2
2
2
2
7
3
10
16
30 32
3
5
40
45
90
67
98
110
120 11
0 1
00
100
100
2
1
1
1
1
1 1
1
1
1
3
1
2
2 3
3
10
10
15
2
2 2
2
2
2
5
15 5
5
5
5
2
8
1
1 1
1
1
1
7
14
37
40
42 44
4
6
48
50
1
Approved For Release 2002/05/13: CIA-RDP78-01092A000100100006-3
 Approved For. Release 2002/05113 : CIA-RDP78-01092A000100100006-3
CHART 1
Page 2
JOB May
June
July
Sept,,;
Dec
Mar
June
Sept
Dec
64
64
64
64
64
65
65
65
65
67
88
81
80
80
80
80
5
2
1
2
3
.1
2
3
1
6
6
6
6
6
2
2
2
2
3
3
4
4
4
3
3
3
3
3
3
1
2
1
2
2
2
2
1
2
2
2
2
2
2
6
6
6-
6
6
6
6
5
1
1
1
1
1
1
1
7
2
2
2
2
18
10
11
25
8
8
8
*Full implementation of monitor with addressable clock will reduce this
time to ?0?. Time will still be valid but distributed among the jobs.
Approved For Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
 Approved For. Release 2002/05/13 : CIA-RDP78-01092A000100100006-3
.. I
CHART 1
Page 3
May
June
July
Sept
Dec
Mar
June
Sept
Dec
64
64
64
64
64
65
65
65
65
1
1
1
1
1
1
10
40
70
90
105
120
130
1
2
2
2
3
3
3
3
1
1
1
1
1
1
1
5
10
15
20.
20
20
20
15
35
35
35
35
35
7
2
2.
2
2
2
5
15
25
35
45
4
4
4
TOTAL
295
309
370
463
540
560
585
621
653
*Applications under study but not yet approved.
Approved For Release 2002/05/13 CIA-RDP78-01092A000100100006-3