by DAN CALLOWAY
Published 21 January 2010 @ 21:06 UTC
WEAVERVILLE, NC – Good morning ladies and gentlemen, my name is Dan Calloway, and I am a consultant for NewTechnologies, Inc. I was hired by your company to propose a new technological development in computer architecture that I presented to the board of directors last week following research that was conducted into your current computer architectural manufacturing process. I have also been asked to chair the team that will be evaluating the proposal that I presented to the board, a proposal that I will be discussing with you today.
First, please allow me to give you some insight into the history of the development of the current computer architecture that most computer platforms are designed around, including ones your company is currently designing. As most of you know, most modern computers are built on the von Neumann architecture and are based on silicon transistor technology. Von Neumann was a Hungarian mathematician who developed the architectural design of the modern computer in papers that he wrote circa 1945. His architectural design is one in which there is a single CPU or central processing unit or chip containing millions of silicon-based transistors, with the CPU containing the control unit and the arithmetic/logic unit that is separated from the memory and Input/Output features. As I stated, most modern computers today are still being built around von Neumann’s architectural design concept even though the original design has taken on several modifications. These modifications have included such things as increasing the word-length within the CPU processor, and the introduction of various degrees of parallelism within the architecture so that one instruction can be processed while another instruction is being fetched simultaneously. In addition, our modern supercomputers are the result of the use of multiple CPUs (still designed around the von Neumann architecture principle), which are used in a parallel computer architectures (Warren, 2004).
Although the architecture and the technology of the silicon-based transistors that make up the CPUs that control our modern computers have been extremely successful over the last 50 years or so, there are problems for which this architecture and technology are not particularly well adapted, and my proposal is to recommend to your company alternatives in both the architectural design and the silicon-based technology (Warren, 2004). For example, from the standpoint of the physical characteristics of silicon itself, there is a point at which continuing to add more transistors per chip (increasing chip density) will begin to violate the laws of physics since further attempts at miniaturization on the chip will become physically impossible from a molecular and atomic perspective. It is the laws of physics that will cause the use of silicon-based transistors to become extinct in the next several years even with the advances in transistor and chip design we have seen over the last decade. Furthermore, from an economic standpoint, the cost associated with continuing to add more silicon transistors per chip will eventually reach a point of diminishing returns where the cost of these additional transistors per chip will not yield sufficient increases in computing power to make them cost effective (Nair, 2002). There are other examples that I could provide, but these are the major ones.
My proposal, therefore, is two-fold: (1) I propose that the current computer architecture of von Neumann has outlived its usefulness in the IT organizations of today and alternatives to the von Neumann design are needed if computing power is to be advanced further in the future, and (2) I propose that the physical and economic limitations of current silicon-based transistor technology that threatens its very existence in the next few years should be replaced with alternative technologies such as biological-, or molecular-based technologies that do not rely on the silicon substrate.
My proposal suggests that advances in current computer architectural design and a solution to replacing silicon-based transistor technology can both be inspired by heuristic solutions found in nature. There are lessons to be learned from how biological systems achieve self-control in a decentralized manner. The natural world provides proof that extremely complex problems in pattern recognition, for example, are solvable as these solutions are observed quite often in the Animal Kingdom. To date, the heuristic approach of finding solutions to complex problems in nature has largely concentrated on implementing these solutions on software running on conventional computers. However, the development of neural networks that mimic the human brain in an attempt to improve instruction execution time has involved developing this technology on silicon-based technologies that will soon become extinct (Warren, 2004).
I suggest the two-fold proposal above is achievable through the implementation of a biological-, or molecular-based solution to computing, which involves the replacement of silicon-based transistor technology with organic or inorganic molecular and biological material through the process of constructing switches using molecular material that exist in the nanometer size range. The protein, bacteriohodopsin, has been heralded as a viable alternative to silicon technology in the emerging computing arena of molecular computing. A detailed introduction to the work on bacteriohodopsin and its use in building a three-dimensional memory prototype can be found in Birge, et al. Furthermore, advances in computing architecture and technology to replace silicon can be found in the use of inorganic substances, such as lithium niobate used in the development of holographic memory, which exceeds the memory capabilities available from the use of bacteriohodopsin. Utilizing either the molecular-, or biological-based solutions as discussed here will solve the current problems associated with silicon-based technologies and provide for increased computing capabilities to solve increasingly complex problems requiring greater computing power than currently available through conventional computing architectures and technologies that your company currently utilize in its manufacturing process (Warren, 2004).
There is obviously more work that needs to be researched in the field of molecular computing and biological computing, to make it work effectively and efficiently but my proposal is a starting point in getting your company to entertain the frontiers of alternative computing architectures and technologies that will revolutionize the IT and computing industry as we know it today (Warren, 2004).
It has been my pleasure to address you today, and I hope that you will allow me to work closely with your CTO and your executive team in developing implementation plans to transform your business processes to begin the transition from silicon-based transistor technology to molecular-, and biological-based technologies that will play a vital role in computing in the future.
References:
Birge, R.R., Gillespie, N.B., Izaguirre, E.W., Kusnetzow, A.,Lawrence, A.F., Singh, D., Wang Song, Q., Schmidt, E., Stuart, J.A., Seetharaman, S., Wise, K.J.: ‘Biomolecular electronics: protein based associative processors and volumetric memories’, J. Phys. Chem. B, 1999, 103, pp. 10746–10766
Nair, R. (2002). Effect of increasing chip density on the evolution of computer architectures. IBM Journal of Research and Development , 46 (2/3), 223-234
Warren, P. (2004). The future of computing – New architectures and new technologies. IEEE Proceedings – Nanobiotechnology , 151 (1), 1-9.
Dan Calloway
Recent Comments