Cellular computing opens a new frontier of engineering that will dominate the technology of the next century. Employing information technology, the future holds promise for the development of means to organize and control biological processes that are just as effective as our current mastery of electrical processes.
In particular, biological cells are self-reproducing chemical factories that are controlled by a program written in the genetic code. Current progress in biology will soon provide us with an understanding of how the code of existing organisms produces their characteristic structure and behavior. As engineers we can take control of this process by inventing codes (and more importantly, by developing automated means for aiding the understanding, construction, and debugging of such codes) to make novel organisms with particular desired properties.
Besides the obvious application of control of biological processes to medicine, we will be able to co-opt biological processes to manufacture novel materials and structures at a molecular scale. The biological world already provides us with a variety of useful and effective mechanisms, such as flagellar motors. If we could co-opt cells to build organized arrays of such motors, with accessible interfaces for power and control, we could see how this could be of engineering significance. Common, biologically available conjugated polymers, such as carotene, can conduct electricity, and can be assembled into active components. If we, as engineers, can acquire mastery of mechanisms of biological differentiation, morphogenesis, and pattern formation, we can use biological entities of our own design as construction agents for building and maintaining complex ultramicroscopic electronic systems. Such systems will have better performance and reliability then technologies based on less precisely controlled chemical processes. Of course, one of the most important products of mass-produced molecular-scale engineering will be extremely compact, efficient, and effective computing mechanisms.
Thus, in spite of the long gate delays in cellular computing mechanisms, the fact that cells can reproduce and organize into precisely arranged and differentiated tissues means that we can use them as the (very slow) agents of molecular-scale manufacturing of macroscopic objects. It is the resulting objects that we desire--they may contain electrical circuitry with picosecond cycle times. The slow biological systems are our machine shops, with proteins as the machine tools, and with DNA as our control tapes.
This work is supported, in part, by the DARPA/ONR Ultrascale Computing Program under contract N00014-96-1-1228 and by the DARPA Embedded Computing Program under contract DABT63-95-C130.