X-Mozilla-Status2: 00000000 Message-ID: <[log in to unmask]> Date: Wed, 24 May 2000 12:52:03 -0400 From: Art Weil <[log in to unmask]> X-Mozilla-Draft-Info: internal/draft; vcard=0; receipt=0; uuencode=0; html=0; linewidth=129 X-Mailer: Mozilla 4.72 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 To: Conchologists of America List <[log in to unmask]> Subject: Re: Salt marsh memory? too strange to pass this up..... References: <000701bfc592$79dd3320$cd8824cf@default> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear Peta;- Being a living thing, doesn't the individual bacteria tend to die? Art peta bethke wrote: > Ok, not your usual Conch-L thread BUT a strange new world ,,,,,, please read on...... > > High-Protein Computers > > Carol Levin > > The promise of binary biology. > Chemistry professor Robert Birge may look as if he's brewing beer in his lab at Syracuse University, but what he's really > doing is recreating the environment of a salt marsh, the ideal incubator for growing bacteria that make a certain protein > that shows great promise as the basis for a radically new kind of memory: protein memory. > > Imagine a storage medium the size of a sugar cube that stores a terabyte of information. As odd as it may seem, this device > will be based not on silicon transistors, but on protein molecules that change their shape when exposed to light, enabling > them to store and transfer massive amounts of data. > > While research and development efforts today focus on shrinking the size of silicon transistors, there's also an effort afoot > to manipulate tiny protein molecules that may turn out to be economical for storing data on a massive scale. According to > Birge, director of the W.M. Keck Center for Molecular Electronics at Syracuse University, "The size of a single logic gate > will approach the size of a molecule by about the year 2030." > > This branch of bioelectronics emerged in the early 1970s when two scientists observed the structural changes that a protein > molecule known as bacteriorhodopsin (which is grown by a bacterium that is found in plentiful supply in salt marshes > throughout the world) exhibited when exposed to light. Later, Soviet scientists recognized the potential of the molecule to > act as a switch with on and off positions, essentially the same binary function that a silicon transistor serves. While > silicon alters its state when a current of electricity excites electrons, a protein changes its shape upon absorbing light. A > laser beam controls the switching in a matrix of memory cells (see diagram). > > So would you feel comfortable entrusting your data to a close relative of a protein shake? While an organic entity seems > susceptible to decay, disease, or a ravenously hungry user, this bacterium and its photosynthetic protein, which thrives > naturally in the harsh environment of salt marshes, is resilient and stable. Reliability is another issue. "Writing is a > piece of cake. It's reading that's harder," says Birge. Errors crop up because of noise from the laser interfering with the > read signal. The goal is to reduce the errors to the level found in silicon. > > While in terms of speed, protein-based memory falls somewhere between a high-speed disk drive and semiconductor memory, the > advantage is in its ability to read data in parallel. Birge anticipates the major impact of bioelectronics will be in the > area of 3-D memory, which stores data in all three dimensions. Magnetic, flash, semiconductor, and optical memory store data > only on the surface. "The approach we're taking is to have true volumetric memory where you would gain a 300-fold improvement > [in density]," says Birge. > > The Syracuse lab has been working on prototypes of protein-based 3-D memory and associative memory (used in neural networks > and artificial intelligence), and is preparing to test memory cards in PCs. The question that hasn't been answered yet is > whether protein memory is economically viable, especially as flash memory improves. > > The first hybrid products that use both semiconductor and protein memory (a system would have, say, 3GB to 24GB of protein > memory and 4MB of fast RAM) could reach consumers in as soon as eight years, Birge says optimistically. > > But these developments depend on a few contingencies. Inexpensive lasers are needed, and some company will have to buy into > the technology. Birge says that three companies are evaluating the technology now. "They range from very, very large down to > startup."