VOLUME 5, NUMBER 1 JANUARY/FEBRUARY 1979
Time
Sharing
Costs
A COST COMPARISON
OF COMMERCIAL
TIME-SHARING SYSTEMS
A special report prepared
for ATSU Members
by Real Decisions Corporation
Page 2
Enclosed For ATSU Members Who Enclosed for ASCU Members Who
Subscribe to the Interactive Computing Directories: Subscribe to the BENCHMARK REPORT:
I. FINANCIAL MODELING LANGUAGES Benchmarks of the Datapoint 1170,
V. DATA BASES AVAILABLE TO USERS including speed tests, “real life” problems,
XXX. REMOTE BATCH SERVICES ease-of-use tests, and user comments.
Interactive Computing is published every other month by the Association of Time-Sharing Users, Inc., P.O. Box 9003, Boulder, CO 80301. Second-class postage paid at Boulder, CO
A COST COMPARISON OF
Preface
COMMERCIAL TIMESHARING SYSTEMS
January, 1979
The Association of Time Sharing Users (ATSU) has commissioned Real Decisions
Corporation (RDC) to develop this special report covering a cost comparison
of the results of running one benchmark-type program on seventeen (17) dif¬
ferent timesharing services. The objective of this report is to demonstrate
the necessity of running benchmarks on prospective vendors' services as part
of the total evaluation effort. RDC regularly runs standardized benchmark-
type programs on Remote Computer Services vendors and periodically publishes
the results. ATSU specifically requested this special report to be focused
on one of RDC's standard benchmark programs. Users are cautioned not to
construe these results as being representative of any particular vendor's
capabilities or cost profile.
Making a decision between vendors of Remote Computing Services comes after
a review of your corporate DP history, current facilities and level of satis¬
faction in line with your basic objectives and long term plans. Once you
understand your current workload, project future requirements and establish
acceptable standards, you are ready to get down to the serious work of vendor
selection or review. There are many elements to consider in determining the
right vendor for you and your application. This report only deals with one
of those elements - benchmark costs. While costs are only one factor to con¬
sider in making a vendor decision, RDC believes that the dramatic differences
in costs documented herein provide enough reason to carefully evaluate alter¬
native vendors.
The vendors contained in this report were chosen by RDC on the basis of its
experience in the marketplace. RDC does not represent that these seventeen
timesharing services comprise an exhaustive list of all such RCS vendors.
Other may choose to evaluate different vendors for reasons pertinent to their
needs.
From time to time, RDC runs its standard benchmark programs on additional RCS
services to expand the list of vendors contained in its reports. Interested
vendors should contact RDC for the criteria necessary to be included in future
reports.
©Copyright 1979 by Real Decisions Corporation.
Vendors included in this Special Report are:
Name
ADP Network Services, Inc.
Boeing Computer Services
Boeing Computer Services
CallData Systems, Inc,
CallData Systems, Inc.
Computer Sciences Corporation
CompuServe Network, Inc.
Control Data Corp. (CYBERNET Services)
General Electric Information Services
McDonnell Douglas Automation Co,
National CSS, Inc.
On-Line Systems, Inc.
Rapidata, Inc.
Service Bureau Company
Tymshare, Inc.
Tymshare, Inc.
United Computing Systems, Inc.
Headquarters Hardware
Abbr.
Location
Utilized
ADP
Ann Arbor, MI
DEC
BCS-C
Morristown, NJ
CDC
BCS-I
Morristown, NJ
IBM
CDS-D
Woodbury, NY
DEC
CDS-H
Woodbury, NY
Honeywell
CSC
El Segundo, CA
Univac
CSV
Columbus, OH
DEC
CYB
Minneapolis, MN
CDC
GE
Bethesda, MD
Honeywell
MCA
St, Louis, MO
CDC
NCS
Wilton, CT
IBM
OLS
Pittsburgh, PA
DEC
RAP
Fairfield, NJ
DEC
SBC
Greenwich, CT
IBM
TYM-D
Cupertino, CA
DEC
TYM-I
Cupertino, CA
IBM
UCS
Kansas City, MO
CDC
PRACTICES, PROCEDURES AND PROGRAM DESCRIPTIONS
RDC arranged directly with the vendors covered by this report to utilize their
services. All runs were performed on 30 cps terminals during the summer and
early Fall of 1978. The price schedules quoted were in effect as of November
1978. Where published price lists have been modified since the runs were made,
but the underlying pricing algorithm remained unchanged, RDC updated the bench¬
mark costs appropriately without re-running.
RDC recognizes that most vendors offer attractive discounts for large users
and long-term commitments. In addition, many vendors offer discounts for
running programs at less than interactive services levels or on non-prime
time schedules. In order to evaluate any discounted or specially-priced
proposals from vendors, however, RDC believes that a user must start by making
cost comparisons using the retail or standard rates that timesharing services
publish. Therefore, all prices used in this report are based upon the stan¬
dard prices for interactive computing during prime time as available for be¬
ginning or low-usage customers.
All program runs were performed by RDC personnel experienced in benchmarking
timesharing vendors. Where more than one valid implementation of the bench¬
mark program's function was possible, RDC sought and employed the recommenda¬
tions of the vendor's technical assistance staff. In sum, great care was exer¬
cised to obtain uniform results for all services.
Run costs shown are for both compilation and execution (including linking and
loading, where necessary) - and are not just for execution of pre-compiled
programs. The run mode resembles program development more than fixed produc¬
tion. To illustrate that CPU costs are only one part of the total costs for
a particular run, arbitrary amounts of connect time and disk storage utiliza¬
tion were added to the run. The following charts and graphs reflect the re¬
sults of adding the costs of 3 minutes of connect time and 1500 characters of
disk storage to the actual CPU costs incurred in running the benchmark program.
The standard BASIC version of the New Product Planning Problem (used in the
RDC report "Financial Modeling Decisions," published December 1978) was selec¬
ted for the purpose of this report. A description of this problem follows:
New Product Planning Problem
w
The market research department of a major company is evaluating the profit¬
ability of a new product over the next four (4) years. A model is created
as follows:
STARTING POINT
Units Sold = 50,000 -
Selling Price = $8.50
Variable Costs per Unit -
-Raw Materials
-Direct Labor
-Packaging
-Distribution
Fixed Costs -
-Factory
-Other
Effective Tax Rate = 22%
MODEL 1 S ACTION
Increase 15% Per Year
Increase $.50 Per Year
Inflation for next 3 years
at 7%, 8%, 6%
$3.00
$ 2.00
$0.50
$0.75
Constant for 4 years
$25,000
$15,000
Constant %
A base line run is established, and several parameters are varied in a "what-if"
mode on subsequent runs. Program output runs two to three pages and is printed
in a standard report format of report line items across column years.
- 5 -
VENDOR COSTS FOR CPU, CONNECT, AND STORAGE
AS OF NOVEMBER, 1978
VENDOR
CPU
CHARGES
CONNECT/HR
AT
30 CPS
STORAGE
(1000 CHAR/
MONTH)
ADP
$.02/CRU
$15.00
$1.00
BCS-C
.20/CCU
8.50
0.33
BCS-I
. 15/CSU
10.00
0.18
CDS-D
.03/Sec.
8.00
0.39
CDS-H
.10/SRU
11.00
0.38
CSC
.33/SRU
12.50*
0.41
CSV
.023/SRU
15.00
0.49
CYB
.35/SBU
15.25*
0.38
GE
.13/CRU
12.75**
0.11
MCA
.18/MRU
10.00
0.19
NCS
.20/ARU
13.00
0.18
OLS
.05/CPU
10.00
0.47
RAP
.07/CPU
13.00
0.60
SBC
.18/PU
15.00
0.66
TYM-D
.13/TRU
13.00
0.45
TYM-I
.27/TRU
13.00
0.14
UCS
.15/SU-B
22.00
0.24
*Net charge based upon 25,000 characters/hour
**Metro Access Option 2 at 25,000 characters/hour
COMPARATIVE COST TABLE
NEW PRODUCT PLANNING PROBLEM
BASIC LANGUAGE
VENDOR
CPU COST
% OF
HIGH
VENDOR
CONNECT
COSTS
(3 Min.)
STORAGE
COSTS
(1500 CHAR)
TOTAL COST
% OF
HIGH
VENDOR
ADP
$.88
55%
$.75
$1.50 7
$3.13
100%
BCS-C
.80
50 2
.43
.50
1.73
55
BCS-I
.41
26
.50
.27
1.18
38
CDS-D
.21
13
.40
.59
1.20
38
CDS-H
.20
13
.55
.57
1.32
42
CSC
.36
23
.63
.62
1.61
51
CSV
.56
35
.75
.74
2.05
65
CYB
.41
26
.76
.57
1.74
56
GE
.71
44
.64
.17
1.52
49
MCA
.34
21
.50
.29
1.13
36
NCS
.72
45
.65
.27
1.64
52
OLS
1.60
100 2
.50
.71
2.81
90
RAP
.93
58
.65
.90
2.48
79
SBC
.36
23
.75
.99
2.10
67
TYM-D
.89
56
.65
.68
2.22
71
TYM-I
1 .20
75
.65
.21
2.06
66
UCS
.18
11
1.10 3
.36
1 .64
52
Selected qualifying notes to illustrate need for broader evaluation:
^ ADP storage volume discounts yield dramatic reductions.
L The costs for these vendors on this program are not consistent with
^ RDC's general cost profiles.
5 UCS price change 1/1/79 will reduce connect charges.
- 7 -
CPU COSTS
NEW PRODUCT PLANNING PROBLEM
VeAczntagu o{) Highest Vmdosi
- 8 -
TOTAL COSTS
NEW PRODUCT PLANNING PROBLEM
V&icmtagz Highest 1 /andoft
- 9 -
CONCLUSIONS
The preceding graphs illustrate quite dramatically the differences in costs
associated with running only one program on seventeen different RCS vendors.
These kinds of differences justify careful evaluation of all pricing elements
when choosing a particular vendor*s service. Note how the cost relationships
of these vendors change when factors such as connect and storage costs are
added to total CPU costs. Vendors who appear high based solely on CPU costs
can dramatically shift on the total cost graph, and vice versa.
In addition, this problem is only one of a series of benchmark type programs
which a user must run in order to gain an accurate cost profile of any par¬
ticular vendor*s service. RDC itself runs many benchmark programs designed
to test various combinations of CPU time and memory, I/O file uses, and com¬
putational speeds. Furthermore, BASIC is only one of the languages used in
timesharing and, in fact, vendors often display entirely different cost re¬
sults for programs coded in FORTRAN versus BASIC.
Users are encouraged to devise additional benchmark programs which will not
only provide more complete vendor cost profiles, but also display ease of use
characteristics relative to the various Command and Editor facilities. RDC
strongly recommends that users perform extensive benchmarking tests which re¬
flect their particular requirements for Remote Computer Services.
ATSU readers are urged to use the results presented in this special report
with caution. The main purpose of this report is to stimulate users to per¬
form extensive benchmarks prior to selecting vendors for their use because
the fact is that costs do vary widely from vendor to vendor.
The value of a service is determined by a combination of factors in the de¬
cision process, weighted by the specific user needs. It is important:
• To understand the PROFILE of a vendor - his business foundation,
emphasis and practices;
• To examine the FACILITIES - that specific set of hardware, software
and network utilized;
• To evaluate the available SUPPORT - types of personnel, applications
and educational help;
• And, with this background, PERFORMANCE can be judged - based on
user references, economy of operation and overall service
capabilities.
In sum, the evaluation of factors described within the specific decision
process will result in the following equation:
PROFILE + FACILITIES + SUPPORT + PERFORMANCE = CHOICE
The processes described above represent some of the analyses users should
perform when deciding among potential vendors of Remote Computer Services.
Note that one specific aspect of PERFORMANCE deals with economy, and one
factor under economy concerns benchmarks - a procedure widely used to test
efficiency, measure response time and evaluate costs of various services.
It is this specific economic factor - benchmarks - which this report ad¬
dresses. Results from such benchmark efforts can contribute to the overall
task of evaluating competing services, but the proper use of benchmarks is
essential. They are certainly not the whole story, and total results must
still determine each user's choice.
EDITORIAL
I’m constantly amazed at the incredible difference in
prices between remote computing suppliers. As the
chart on page 7 shows, a short test program that
costs $1.13 on one service costs $3.13 on another.
But what about the results of other test programs, or
better yet, larger “real life” programs? The amazing
fact is that the results change dramatically as
different types of programs are run on each sup¬
plier’s system. Referring to RDC’s 124 page report
entitled “The 1978-1979 Comparative Time-Sharing
Cost Analysis Report,” (available for $695 from
RDC) we find that ADP Network Services and On-
Line Systems — both of whom had deceptively high
prices on page 7 — each had very low prices for
other types of programs. This seems to be telling us
two things: First of all, that we should not jump to
conclusions about any particular supplier based
upon one example of someone else’s benchmark
program, and secondly, that what really counts is
how the suppliers compare with one another using
your own applications programs. Clearly, this study
all by itself cannot be the sole factor considered in
making an intelligent purchase decision. But it again
demonstrates the incredible differences between
remote computing systems and services — differ¬
ences which are important for us as users to
\ _, understand and appreciate. HS
ATSU’S CORPORATE ASSOCIATE MEMBERS
ADP Network Services
American Terminal Leasing
Avco Computer Services
Boeing Computer Services Company
CallData Systems
Citibank - Interactive Computing Center
Corporate Time-Sharing Services, inc.
Datanetwork, Honeywell, Inc.
Data Resources, Inc.
General Electric Company, Information Services
Informatics - Data Services Div.
Insco Systems Corporation
Interactive Market Systems, Inc.
I. P. Sharp Associates, Ltd.
Litton Computer Services
Metrocom Inc.
Minicomputer Modeling, Inc.
National Computer Network
On-Line Systems, Inc.
Quantum Science Corporation
Rapidata, Inc.
Scientific Time-Sharing Corporation
SDC Search Service
Sun Information Services
Telenet Communications Corporation
Tentime Company
Time-Sharing Resources, Inc.
Trendata
United Computing Systems
University Computing Company
Vocal Interface
Warner Computer Systems
Western Union - Data Services Company
Zeta Research
- 11 -
ATSU and ASCII Chapters, Local Contacts and Special Interest Contacts
ALABAMA
Ray F. Heyd
Birmingham — ATSU Local Contact
Vulcan Materials Company
(205) 877-3000
ARKANSAS
Gene Dugger
Searcy — ASCU Local Contact
Harding College
(501) 268-6161
CALIFORNIA
Richard Dumas
Mountain View — ATSU Local Contact
Commodity Research Institute
(415) 941-4646
Gary Galan
Newport Beach — ATSU Local Contact
Commercial Bankers Life Insurance
(714) 833-8450
Frederick Gallegos
Los Angeles — ATSU Local Contact
U.S. Gen’l Accounting Office
(213) 688-3809
Don Hatch
San Diego — ASCU Local Contact
Christian Mgmt. Consulting Services
(714) 293-3200
Frank Slaton
San Bernardino — ATSU Local Contact
California State College
(714) 887-7293
CONNECTICUT
Frank Chew
Greenwich — ATSU Local Contact
Amax, Inc.
(203) 622-2824
Charles J. Clock, Jr.
Special Interest Contact for
Educational Applications
West Hartford Public Schools
(203) 236-6081
FLORIDA
William A. Rousseau
Pompano Beach — ATSU Local Contact
Alpine Engineered Products, Inc.
(305) 781-3333
J. L. VanGoethem
Miami — ASCU Local Contact
Ryder System, Inc.
(305) 593-3726
IDAHO
Rick Simon
Boise — ATSU Local Contact
Morrison-Knudsen Company
(208) 345-5000
ILLINOIS
♦ Leon Stevens
Chicago - ATSU Chapter President
Standard Oil Company
(312) 856-6689
John A. Koziol
Chicago — ATSU & ASCU Local Contact
Continental Materials Corp.
(312) 565-0100
KENTUCKY
Clyde Jenkins
Special Interest Contact for APL
Humana Inc.
(502) 589-3790
LOUISIANA
W. D. Landry
Abbeville — ASCU Local Contact
Coastal Chemical Co., Inc.
(504) 893-3862
MARYLAND
R. G. Korbeck
Baltimore — ATSU Local Contact
Baltimore Gas and Electric Company
(301) 234-6687
METRO WASHINGTON, DC.
Frank E. Rockwell
Glen Dale — ATSU Local Contact
Astro Data Systems
(301) 982-5996
A. Steven Wolf
DC — ATSU Local Contact
U.S. General Accounting Office
(202) 655-4000
MICHIGAN
J. Ben Friberg
Grand Rapids — ATSU Local Contact
Rapidstan Inc.
(616) 451-6682
Tom Hunt
Cadillac — ATSU Local Contact
Kysor Industrial Corp.
(616) 775-4646
♦ Larry Leslie
Special Interest Contact for
Time-Sharing Administrators
Upjohn Company
(616) 323-4000
MASSACHUSETTS
♦ Stuart Lipoff
Boston — ATSU Local Contact and
Special Interest Contact for
Software Standards
Arthur D. Little, Inc.
(617) 864-5770
MINNESOTA
L. R. Bakewell
St. Paul — ASCU Local Contact
Real Estate Dynamics, Inc.
(612) 698-8891
MISSOURI
Dann E. Kroeger
Kansas City — ASCU Local Contact
Townsend Communications, Inc.
(816) 454-9660
NEBRASKA
Doug Goldsmith
Omaha — ATSU Local Contact
Omaha Public Power District
(402) 536-4015
NEW JERSEY
Jim Fitzpatrick
Special Interest Contact for
Data Base Applications
American Broadcasting Corp.
(201) 488-2345
Robert J. Loring
Haddonfield — ASCU Local Contact
Cardiac Long-Term Monitoring SVC
(609) 795-2220
♦ Bennett Meyer
Special Interest Contact for
Data Security
Singer-Kearfott
(201) 256-4000
Robert Pickford
Northern N.J. — ATSU Local Contact
Warner-Lambert
(201) 540-2999
NEW YORK
Dr. Dina Bedi
Special Interest Contact for
Educational Applications
Baruch College
(212) 725-3196
Terri Gendron
Briarcliff Manor — ATSU Local Contact
Phillips Laboratories
(914) 762-0300
Samuel Leonard
Elmira — ATSU Local Contact
Thatcher Glass Mfg. Co.
(607) 737-3459
Stephen Mandell ONTARIO
New York City - ASCU Local Contact * David Wilson
Citibank N A Toronto — ATSU Local Contact
(212) 559-6242 P S - Ross & Partners
(416) 363-8281
Philip N. Sussman
New York City — ATSU Local Contact
International Paper Company
(212) 490-5827
NEW YORK CITY CHAPTER
Executive Board:
Aram Bedrosian
TWA
Bion Bierer
Bristol Myers
Victor Bittman
Chase Manhattan
Charles Browning
Phelps Dodge
Dennis Callahan
Goldman Sachs & Co.
Chester Frankfeldt
Continental Group
Carl Heimowitz
Harcourt Brace Jovanovich
Alan Kornbluth
American Express
Susan McCain
Morgan Guaranty
Indira Singh
Salomon Brothers
Philip Sussman
International Paper Co.
OHIO
Dennis Bender
Cincinnati — ATSU Local Contact
Procter & Gamble
(513) 562-2469
Ed Casper
Cleveland — ATSU Chapter President
Diamond Shamrock Corp.
(216) 694-3366
♦ Howard Tureff
Cleveland — ATSU Local Contact
Diamond Shamrock Corp.
(216) 694-5963
OREGON
Roland S. Hanson
Portland — ATSU Local Contact
Oregon Assoc, of Hospitals
(503) 228-5608
PENNSYLVANIA
* Dale Hummer
Pittsburgh — ATSU Local Contact
Westinghouse Electric Corp.
(412) 256-4889
Steven D. Rellis
North Wales — ATSU Local Contact
Leeds & Northrup Corp.
(215) 643-2000
D. T. Wu
Philadelphia — ATSU Local Contact
DuPont De Nemours & Co.
(215) 339-6307
TEXAS
Ralph N. Bussard
Houston — ATSU & ASCU Local Contact
Price Waterhouse & Company
(713) 654-4100
WISCONSIN
Anil K. Bhala
Green Bay — ASCU Local Contact
L. D. Schreiber Cheese Co.
(414) 437-7601
David J. Ritter
LaCrosse — ASCU Local Contact
LaCrosse Garment Mfg. Co.
(608) 785-1400
John J. Stewart
Wausau — ASCU Local Contact
Van Ert Electric Co., Inc.
(715) 845-4308
Paul Thoppil
Milwaukee — ATSU Local Contact
RTE Corporation
(414) 547-1251
Become an ATSU or ASCU Local Contactforyourarea. Your name
and telephone number will be listed on this page in each issue of
Interactive Computing, enabling other members to contact you
with their questions. Only users, not suppliers, areeligibleto apply
by writing to the Association.
ATSU and ASCU Council Members.
Published jointly by the Association of Time-Sharing Users and
the Association of Small Computer Users. ©Copyright 1979, P.O.
Box 9003, Boulder, Colorado 80301. Telephone (303) 499-1722.
Leon Stevens *■
Vice President
Hillel Segal *
President,
Publications Editor
Martin Neville
Secretary
Earl Carroll
Treasurer
ATSU is an independent non-profit association providing a forum
for the discussion of remote computing topics. ASCU, organized
as a sister association to ATSU, is also independent and non¬
profit and is devoted to serving the informational needs of small
computer users.
