MEMORIES – TECHNOLOGY AND MARKETS FOR THE 70’S
Technical Judgements: Business Decisions
I would preface my comments with the acknowledgement that forecasts of change must weigh carefully the inertia of opinion.
Memories for computers will represent the largest dollar volume for an identifiable product entity in the electronics industry for the 70’s. Early in the decade, monolithic semiconductors will become the dominant main memory technology representing m~re than half of the new sales by 1975 and constituting the majority of in-use main memory before the close of the decade. Read only memories will be an insignificant percentage of the total memory market, due principally to diminishing cost performance advantage, high tooling and support cost and high potential obsolescence.
There are a number of compelling reasons for this to occur – most significantly:
Semiconductor memories offer raw performance in terms of speed and density which is unachievable in other current technologies.
With sub i00 nanosecond memories now in production and new designs having another order of magnitude increase in speed, now in the laboratory, semiconductor memories constitute an exclusive performance class.
Monolithics, being a homogeneous technology, offers the potential of entirely new system architectures. These new architectures will involve parallelism to a degree where today’s definition of memory and processor will seem nonsensical. Hard implementation of memory hierarchy will provide transparency for all levels of programming. Associative memories will become econonkically feasible and it can be expected that associative concepts will be extended to include synoptics.
These new architectures can be expected to contribute substantially toward the solving of the problem of software economics.
Being a process-based technology, monolithics have an intrinsic reliability unachievable by technologies involving a piece part assembly base. Further, where the assembly-based technologies have reached, or are rapidly approaching, a plateau in terms of mechanization and will thus have their manufacturing costs directly tied to labor, monolithics have the’potential for significant increases in densities, yields and packaging innovations.
Viewed from a different perspective, one which encompasses overall computer technologies for the 70’s and beyond, monolithics must become the dominant technology if for no other reason than simply that the rate of expenditures within the industry has become so high that we cannot now afford a transitional technology having a known obsolescence. While the inertia of capital investment and engineering and manufacturing skills will be a pressure to extend the life of magnetic technology into the 70’s, the threat of early competitive impact of systems utilizing magnetics will cause monolithics to prevail.
Even the now promising bubble may burst under the competitive pressure of investment in semiconductor technologies with array densitites approaching one million bits per square inch.
To this date, to a great extent, the’management of our industry could be characterized as a tug of war between the engineer and the marketeer. While this was probably a necessary part of the development period of our industry, it is now a luxury we can no longer afford. We have reached a level of business where we can no longer afford to innovate for the hell of it or to market on a basis solely of~ faster, more exotic device. )
With the change to principally a process-based technology, one characterized by very high capital investment and a heavy scientific supporting base, many companies must necessarily re-examine their proprietary posture of what and from whom they buy and what and to whom they sell. To those companies who continue to plan their business solely to maximize value added in the p~oduction of hardware, great trauma lies in store.
Where the engineer, the programmer and the marketeer previously made decisions, they will now provide judgement contributing to a more broadly based business decision. Technologies having reached a plateau that is sufficiently broad to discourage the expenditures necessary to attempt the scaling of the next precipice, efforts will now be concentrated onthe refinement of technologies and broadening of the market base.
Submitted by George R. Cogar President Cougar Corporation Herkimer, New York
IEEE – International Computer Group Conference & Exposition Washington, D. C. – June 16-18, 1970.
The 70’s will see a new era of ent~epreneu~ship. Opportunities wi~ increaae in line with the growth of new technologies.
The important ventures of the 70’s, however, will differ significantly from those of the 5O’s and 60’s, in that the ventures of the 7O’s will sta~ on a larger scale.
New technologies are more complex and will require a broader spectrum of disciplines to form a viable organization with sufficient freedom of choice in the planning of it~ development and growth.
The o~ook of people comprising the knowledge industries ~ expanding beyond the perimeter of their own individual specialities.
They are searching for challenges that will also involve the wtilization and management of new idea~ and new technologies.
Knowledge ~ the new wealth, the new ea~, and the people who are making the greatest contribution to tt~ compound growth are going to in6ist on having a proportionate voice in its direction and u~e.