United States. Congress. Senate. Committee on the.

The Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) online

. (page 60 of 140)
Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 60 of 140)
Font size
QR-code for this ebook


Mr. O'Leary. Can you give us an example of a user to whom this
kind of service would be especiallv appealing?

Mr. Saxders. AVell, I think we make all the Avis Wizard terminals,
for example, that you see in the different airports. Xow, there's about
a thousand such terminals, and they can only speak because of the
wav the telephone network would be, or the nonswitched system that
IBM would talk about, to a pair of 360/G5's in Garden City.

Now, in going to rent a car at the terminal, the girl takes your ap-
plication for the car and she makes up what is known as a transaction
wliich the telephone system sends to Garden City, it is then processed
by the computer and sent back.

The example I'm giving is those 3G0/G5*s, which cost about $5 mil-
lion a year to operate and are about 26 percent utilized.

Now, in a switched system you would be able to, let's say, access and
make your hotel reservation; you might be able to pick up your air-
line ticket; ever\^thin^ would be done at one particular point, with
a great deal of convenience for you as the user, because the same ter-
minal, as such, which is a fairly good example of a transaction, termi-
nal — designed maybe 4 or 5 years ago, but it's still a pretty good ex-
ample of a transaction terminal — would be effective in this area.

And, the illustration of just making the one reservation versus, let's
say, lining up your entire trip would be a good one.

Mr. O'Le-vry. The system that you describe for Avis Rent-a-Car
now presentlv o-oes through telephone lines to reach the computei'S in
Garden City"? ^

Mr. Saxders. That's correct. In either system you would locate a ter-
minal on the rooftop of the building in which Avis houses its com-
puter and it would bvpass all of the A.T. & T. network that leads up
to that.

There would be an antenna on the airport that would connect up
with all the terminals that happen to be located in the airport. So,
vou see, in tliat way you would bypass the use of all of A.T. & T/s
lines.



5440

Mr. O'Leary. What differences, if any, would exist with respect to
cost and speed ?

Mr. Sanders. I think the minimum saving, for example, between
the Avis case — I think the Avis telephone bill is arovmd $11/^ million
a year — and the saving in the telephone bill would be a factor of two
or three times.

Their computer utilization — because if they were connected up to
it, they could sell the transaction calculation capability elsewhere —
would again be reduced by a factor of two or three times.

It could probably run about 75 percent capacity instead of 26
percent capacity. These total savings, that might result to Avis about
$3 or $4 million a year, could be then passed on to the customer in
just that one accomit. And that is just one typical account that would
be thus affected.

Now, unless you really make this computer capability available,
and useful, by means of a switched system, if you are in IB^M's posi-
tion you would tend to regulate the amount that would become avail-
able and only introduce the switched system much later in time,
because otherwise you would go through another period like you did
in 1970 and 1971, except probably a great deal more severe, when we
had an oversupplv of main-frame capacity.

I think this probably explains why IBINI is taking the risk of getting
so close to being regulated by going into the communications market.
Mr. OXeary. Your statement describes the problems you had, if I
recall correctly, with the IBM 3270, and when you say your terminal
was one with intelligence, I assume that means it can perform a com-
puting function, it has a memory and some logic ?

Mr. Sanders. Well, I think Avhen we talk about intelligence, I thmk
most anything but the simplest terminals, including our earliest ter-
minals, had intelligence.

The earliest terminals which we put out about 1966 had as their
feature the fact that they had more intelligence than our competitors.
What happened with the 3270— and this is borne out by the IBM
management-Telex information— was that IBM determined that we
had a much better terminal, and so they copied our terminal. We had
a clever way of putting that intelligence in the terminal by using a
thing call ed "attribute characters."

So, when they copied this — our design — and we hoard this, we
thought this was a very — you know, imitiation is the most sincere form
of flattery, except when they came out with the 3270, which was
basically a modern copy of "our earlier 720 terminal, they used at-
tribute characters, but they used different attribute characters. And
as a net result, when they introduced them and changed the software,
it was just like changing the track gage.

Until they reduced the software support, in fact, our sales went up.
But as soon as they employed the anticompetitive action of changing
the software, then everybody said this is the game IBM's going to play,
and in 1 month our rate of sales went down from a booking rate of
$3 million worth of terminals a month down to $1 million worth of
terminal? a month ; and stayed there.

It tools: a very, very long time and a complete change of line before
it built up again.

Mr. O'Leary. I want to see if I understaud it and can make it clear
for the record. Your product requires software and your software
has to be able to interface or be compatible with their software.



5441

Mr. Sanders. Yes.

Mr. 0'Le.vky. When they say they will not support that software,
putting this in simple terms or crude terms, does that mean that some-
one from IB]\I does not come out and update or rewrite that software
which they are using, which has to be compatible with yours?

Mr. Sanders. Well, the software that the customer was using will
work, and continue to work. What the effect of the software with-
drawal means is that IBJNI will no longer make its improvements com-
patible, and it's an effective disconnect because the customer has built
up a situation of dependence on this. So by withdrawal of that, effec-
tively the customer knows well, I've got 4 months, I've got 6 months;
depending upon my particular situation, maybe I can stand still 9
months.

But he makes the decision to terminate and he does terminate. And
you get millions of dollars of terminals that then slowly start to flow
back in because he has disconnected you, very arbitrarily, in the matter
of the software.

Now, when we put pressure on IBM a year later they finally rein-
stated that software support. But it has not been effective in terms of
really substantially canceling the flow because the trend and the switch
to the other machines was then clearly in progress.

Mr. O'Leary. Does this problem also exist where you are dealing
with terminals that do not have intelligence or do not have the same
intelligence as the product line that you described ?

Mr. Sanders. Well, it is not the software switch and the software
disconnect has not yet happened on all the terminals. In remote job
entry, which is a very popular type of situation, the switch has not
occurred.

We are positive that the switch is coming and that these arbitrary,
capricious types of switchouts are occurring. And we believe it is one
of the fundamental domination policies of IBM.

We have negotiated as hard with IBM as we know how, and they
have told us that if we want a change, that we must go to court.

Mr. O'Leary. Now, you fear, via the acquisition of CML, they will
be able to offer the kind of service that you described with respect to
Avis, and that because of the fear of lack of support for switching the
software interface, a terminal manufacturer such as yourself would be
effectively frozen out ?

Mr. Sanders. Yes. There's a good example in the type of communi-
cation interface called SDLC. This is a very good example, and very
pertinent, and one that they are in the process of using.

SDLC is a new communication interface that is almost designed for
remote job entry, or long transmissions of data. There was a standard
communication interface that was designed by the industry. It was
called ADCCP. It was a much more effective interface and the — all of
industry was working with Mr. McCloskey's CBEMA — to arrive at a
good industry interface.

But like every other standard that industry has tried to develop, at
the last minute IBM switches to one it's been developing on its own.
This is SDLC. ^

_ It's almost like the standard, but not enough. Now in order to sur-
vive, every manufacturer must know what the SDLC interface is.
So there is an interface card that is designed, and you design that



5442

interface card rif^lit up to the point of the last bit of information
that is yet to be released.

The exact last detail that can be arbitrarily switched is yet to be
released so we have a little section of the card that is yet to be designed.
B}' having the design in that stage we are able to plug in the com-
ponents that are missing when IBJNI finally decides to do this.

Now I have recently found out that all these details tliat we are
missing, they have been teaching to their service people for a period
of at least 12 months. And yet they absolutely I'efuse. and we have
sul^mitted this over mam' legal requests to them, to supply that in-
formation to us.

Tliis is what T call capricious — the capricious management of this
interface — for their advantage.

Mr. O'Leary. Once they release it, how long will it take you to
catch up ?

Mr. Saxders. I believe, by a great deal of management, v;e can
catch up in a few months. In the case of the 3270 it took the people 2
years to catch up witli the interface switch. It dei)ends, really, I think
on how well we're able to do this. In the past, when we weren't expect-
ing it. it took us a long time. I think all the manufacturers now
Avill tend to minimize that time. But even with a great deal of special
management attention, it is still very much of an unfair advantage.

]Mr. O'Leary. In 1056 the Justice Department sought to curb
IBM's power in tabulating machines and they managed to make a
quantu.m leap in data processing. I take it now 3-ou see the same
thing occurring by virtue of this acquisition ?

'Slv. Saxders. I think that tlie brealiup and the action on IBM needs
to be aimed at the way the industry is going to be in the future. For
example.. I would like nothing I say to take away from the excellent
recommendations that v\-ere just made with respect to the component
industry.

But it is important to, I think, understand the way tliat the in-
dustry is going to be in the future, and to keep IBaI fi-om giving up
the thing that is becoming obsolete in the marketplace to going on
to grabluug what is just beginning to emerge.

And this is what we are afraid is happening.

'Mr. O'Leary. Thank you. I have no further questions, Mr. Chairman.

Senator Hart. ]Mr. Chttmbris?

Mr. CiiUMBRis. 'AVe have no questions, Mr. Chairman.

Senator Hart. Thank you, ]Mr. Sanders. I hope you can n^ake that
ap])ointment. Let us take a 5-minute recess.

[Whereupon, a brief recess was taken.]

[The following Avas received for the record :]

MATERIAL RELATING TO THE TESTIMONY OF
ROYDEN C. SANDERS. ,TR.

Exhibit 1. — Pr< i)arc(l Sfafciiioit of Mr. f^dnrJer-s

1. IXTROnUCTION

Sanders Associates views witli concern the announcement made -Tiily .3, 1974
that IB:\I plans to enter the domestic satellite business through acQuisition of
55% of CML Satellite Corporation.

CML has been approved I'v the Federal Communications Commission to start
domestic satellite common carrier services. IBM, with its great financial



5443

resources and marketing power, can, through the CML acquisition, achieve
effective control of the integrated business systems marliet, a market more than
double IBM's current market, as well as a large portion of domestic broadband
c-<)mmunieations. This statement will demonstrate how this will be accomplished.

The evolution of the Information Handling Age, in which virtually every seg-
ment of society has both the need and the means to grapple with the information
<^xplosion, has occuri'ed as a result of the confluence of electronics and American-
.style marketing. No single organization epitomizes both the positive and the
negative aspects of the confluence better than IBM.

Electronics is the common denominator for the media of information and for
tlie devices which translate those media and their messages into useful human
knowledge. The force of marketing has caused old methods and institutions to
l>ecome committed to the new electronics way. The transformation is here, and
rhe marketing power is in place.

The one link missing in the merger of disciplines necessary to forge the com-
lilete information handling instrument was emphatically supplied on July 3, a
sleepy afternoon in AA'ashington, the eve of a four-day weekend, with a govern-
ment focused on many divergent problems and the FCC in the throes of reorga-
nization : IBM announced its intention to enter the common carrier communi-
cations field via the acquisition of a controlling interest in CML.

In that one simple stroke, the entire information handling market was
encapsulated.

Satellite communications systems have the potential of becoming the least
■expensive method of communications between computers, remote terminals, and
other business devices. Although IBM will not control the only satellite com-
munications system, it alone will possess the resources to control the .$110 billion
domestic market that encompasses business equipment, data processing and
business communications — the total market for integrated business systems.

This market is made up of systems and products which collect, process, store
and disseminate business information. IB:m already has a monopoly in the data
processing industry. If IBM controls business-related communications, it will be
in a position, under cover of the complexities of satellite communications tech-
111. lories, to control the interaction of all elements attached to the system. Compet-
itive business systems manufacturers and independent terminal suppliers will
lie forced to conform to IBM "standards". These standards will be disclosed in
the time frame IBM desires, when IBM has a product line designed and deliver-
alile, while its competitors, without equal knowledge of the IBM standards, can
(in]y then l)egin to develop competitive products.

IBM has demonstrated an outright monopolistic intent in the peripherals and
Terminal fields, by capriciously manipulating interfaces and device support tech-
niques. It must be prevented from similar activities if communicatons are added
I<> its marketing resources.

This ]>aper does not treat the ramifications for foreign commerce brought on
by IBM's latest move. The domestic problem is sufficiently severe to warrant con-
centrated attention at this time, but the foreign aspect must eventually be
exi>!ored.

In this paper, some of IBM's practices and their anticompetitive effects in the
data processing world are described. The emerging business systems communi-
carions market is shown to present opportunities for predatory practices that
are parallel in nature l)Ut much greater in scope. The paper concludes with
rw.iiumendations for avoiding those effects.

IT. THE EMERGING INTEGRATED BUSINESS SYSTEMS MARKET

IBM's origins in information handling stem from its early occupancy and dom-
inarion of the electromechanical card-taliulating market. A 19.34 conviction, up-
held liy the TT.S. Supreme Court in 1936, for monopolization and anticomi>etitive
jiractices in the punch card machine field did not deter it from achieving this
domination. Likewise, .Tustiee Department litigation, commenced in 19.52 leading
l'> a 19.56 antitrust consent order relating basically to its tab card activities, did
nnt impede its switch into the next generation of information handling — elec-
trnnic data processing.

The progression into successive generations of EDP systems has not been
without minor antitrust casualties and nuisances for IBM. The premature an-
nouncement of system 360 in 1964 ultimately precipitated private litigation by
I'ontrol Data Corporation. Telex, and others, which has so far resulted in the
dropping of less than $500 million in damages — about four months' monopoly



5444

profits by some estimates — and some uow hollow restrictions. The Justice Depart-
ment, in 1969, instituted yet another antitrust action and is hard-pressed to
begin trial this fall.

The setting is ripe for another settlement which will be next to meaning-
less once IBM is underway on its new tack — total information handling sys-
tems from inception through communications to reception and back, with
processing interjected at steps along the way. If the final judgement in the
Government litigation ignores the portent of this latest adroit, but familiar
swing, many new and modified markets will be foreclosed in their infancy.

The dollar volume for the integrated business systems communications mar-
ket in the U.S. alone in 1975 would add the staggering total of $20 billion in
straight communications services, and about $10 billion in the substitution of
electronic message techniques for traditional hard copy deliveries such as mail,
to the current data processing base.

Data processing, communications, and "mail" currently interrelate within net-
works of computer systems. The next major phase in the evolution of data
processing will be distributed systems in which networks of central computers
and intelligent terminals will require an extensive communications capability.
The principal deterrents to more rapid growth of distributed processing have
been expensive long distance communications and IBM's unwillingness to push
this capability until they had the technology and strategy to control the terminal
market. A satellite system will give IBM the missing ingredients, low cost
communications and the means to absolute market control.

With the development of large networks, terminals submarkets are the fastest
growing parts of the data processing market. With IBM's strategy of mov-
ing from relatively simple hardware interfaces to the use of software to con-
trol device support, IBM has the power to effectively disable all devices man-
ufactured by independent companies.

Communications has been controlled domestically for many years by AT&T.
Clearly the FCC has worked diligently to harness this giant. The FCC has al-
lowed competitive entrants to offer specialized communications, especially in
data and facsimile, on an inter-modal basis, using packet-switching, microwave,
and satellites. Before July 3, 1974, however, no entrant had sufiicient capital or
marketing base to move aggressively forward on all fronts. IBM has the capital
and marketing base and has been proven very aggressive.

If the 1969-1975 growth rate for total integrated business systems, including
communications, is projected to 1980, a total market of $250 billion is estimated.
This is a very worthy target for a company which has vowed to maintain domi-
nance in information handling, and now is equipped to serve every major
segment.

IBM will attempt to capture the market in spite of the current government anti-
trust case, or perhaps as a result of it. A familiar pattern could be followed.
Portions of IBM's current business, obsoleted by the shift in technology towards
distributed processing, could be abandoned in favor of the new endeavor of
communications. Promises to beha?e in its "traditional" market could be ex-
changed for permission to compete in a new market. The marketing base
would not be dissipated, however, and the fundamental pressures would not be
relieved.

III. THE COMPUTER ENVIRONMENT

An examination of Sanders' experience in one submarket of the computer
industry will illustrate what could happen in the integrated business systems
market if the IBM/CML merger is allowed.

The requirements for distributed processing systems involving vast numbers of
terminals is growing rapidly. At present each computer terminal is connected
basically to its own computer. This is similar to the situation in the telephone
industry years ago when each company was a private exchange, not connected
to any other, and, as a consequence, of very limited use. Soon, however, terminals
will be required to selectively address other business equipment networks, and
new services will evolve which will be fantastically useful and extremely
inexpensive.

In order to allow interconnection of computer terminals and other business
systems there must be, in addition to low cost communications, meaningful device
support standards or interfaces agreed to and adhered to by IBM and others.
Further, it is absolutely necessary that this come about in a timely manner.

Until the mid-60's in the computer market, the submarkets for terminals and
for the peripheral devices were quite small. IMB's major growth prospects were



5445

^virh computers. Sanders Associates aud other independent terminal and peri-
pheral equipment suppliers entered into competition with IBM in the sub-
markets. Market penetration was relatively swift, and many innovations were
introduced by the independents. ., , ^ . i ttjtvt

This peuertation and the emergence of distributed processing caused IBM
to turn the full force of its predatory marketing practices on the peripheral
iind terminal submarkets. Without technical superiority, only anticompetitive
moves by IBM could continue its dominance.

Sanders' experience illustrates those tactics, and what could happen in the
integrated business systems market. ^ i, ^

Sanders Associates builds display terminals, using keyboards aud cathode-
rav tubes, to allow human communication and interaction with data processing
systems. These terminals can be used in a conversational, or interactive, mode,
or to feed or accept data in a batch process. ., , ^^

Most Sanders terminals were connected to IBIVI computers, because the latter
<:aturate the market. The interface for all interactive display terminals on
IBM 360/370 computers was the 2260 software and hardware interface. The
Sanders terminals were acknowledged as superior to IBM's terminal product
line by IBM's top management; in the more sophisticated segments of the
industry, most technological innovation has come from independents like Sanders.

To maintain its role as an innovative leader, Sanders announced its new 800
line of intelligent displays designed for distributed processing in mid-1971. Other
independents announced similar displays. At about the same time, IBM an-
nounced its new 3270 display which was essentially similar to the Sanders
terminals introduced in 1965. The 3270 incorporated many features of earlier
products, but was not an intelligent terminal and seemingly presented no real
threat to the independents. IBM began deliveries of the 3270 in the fall of 1972.
Sanders' sales and order trends continued to be favorable.

In the fall of 1972, IBM announced virtual systems software in connection
with its new 370 computer systems. Early in 1973. IBM let it be known that the
2260 device support, through which many of the independents' terminals work,
would not be "supported" under the new virtual software.

This announcement meant that users of the IBM 2260 displays, and com-
petitive displays, including Sanders', operating through the 2260 interface,
could not optimally utilize the new virtual systems IBM 370 computers. Users
either had to drop perfectly adequate, older terminals or recently acquired
intelligent terminals or forego tlie advantages of virtual systems software.

The independent interactive 2260-compatible terminal market withered almost
overnight as a result of this capricious manipulation of device support by IBM,
■despite the superior performance and lower cost of the independents' displays.

IBM apparently concluded that it could not attain its goal of increasing
its share of the display market with its new 3270 display and simply changed
the track gauge.

IBM's display strategy was successful in a related aspect. IBM refuses to
disclo.'e interface specifications at the time of a new product announcement
and, in fact, withholds the information until it makes its first shipments.
Competitors were not able to deliver 3270 compatible displays until early 1974,
almost 24 months after IBM's initial 3270 shipments, (the length of time
required for design and engineering) and almost three years after IBM's 3270
product announcement and associated order taking.



Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 60 of 140)