DNA scheme builds computers

By Kimberly Patch, Technology Research News

A decade or so down the road the steady improvements in computer processing speed that we've grown used to are likely to bump up against the laws of physics. The lithography used to make the transistors that form computer circuits uses light and chemicals to etch features into silicon; at some point the process will not be able to form smaller transistors. Most speed improvements in computer processors, however, are due to smaller transistors.

One new way to make smaller transistors involves the stuff of life -- DNA.

Researchers from Duke University, Rambus Inc. and the University of North Carolina at Chapel Hill have devised a pair of computer architectures that would be built from self-assembling DNA.

Computers assembled by DNA have the potential to be extremely small, fast and inexpensive, and would consume very little power.

A single strand of DNA is made up of four bases -- adenine, cytosine, guanine and thymine -- attached to a sugar-phosphate backbone. Adenine and thymine bases, and cytosine and guanine bases have the ability to pair up. Two single strands of DNA whose base orders allow them to completely pair up form the familiar double helix of biological DNA.

Researchers have tapped the self-assembly ability of DNA by producing strands of artificial DNA that have segments of base pairs that connect together in certain patterns. Previous research has shown that it is possible to coax DNA to self-assemble into three-dimensional structures. DNA can also be engineered to attach to other materials in order to include those materials in the self-assembly process.

The researchers' architectures call for single-stranded artificial DNA molecules that have silicon nanorods attached to their ends to assemble into circuit patterns. The DNA junctions between rods are then plated with metal to form the circuitry.

Because the self-assembly process is error prone, the system design accommodates expected errors by building redundancy into circuitry, said Dwyer.

The key is to design a system with a large number of simple components so that any one circuit failure will have a minimal effect on the whole system. The architectures involve as many as one trillion small processors working in parallel.

The researchers' approach is distinct from DNA computing research, which taps the ability of DNA to self-assemble to form strand arrangements that represent computations. In contrast, the researchers' DNA architectures harness DNA self-assembly to form nanoelectronic circuits.

The challenge in designing the architectures was designing circuitry that could self-assemble and devising ways to evaluate the entire system, said Dwyer. "This work introduces the tools to do this along with two self-assembled computer architectures," he said.

What the researchers came up with as viable architectures look like relics from the early infancy of computing, said Dwyer. "Architectures like these haven't been used for ages," he said.

The architectures follow decoupled array-multiplexer (DAMP) and oracle designs, respectively. These are simple parallel processing architectures that could be used to process large optimization problems like the classic problem of routing a traveling salesman through multiple cities. This problem requires a lot of computer power because the possibilities rise exponentially with each added city; there are billions of possible routes through 15 cities, for example.

The researchers' decoupled array multi-processor architecture involves processors that communicate only through a central control unit.

The oracle architecture involves a computer assembled for a particular problem using DNA to match question-answer pairs. The oracle-based computer responds to a query with an answer if the query contains the answer's question.

The ultimate goal is a computing infrastructure that far exceeds today's capabilities in cost, speed, size and power consumption, said Dwyer. "Self-assembly can do this," he said.

It will be 5 to 10 years before a proof-of-concept self-assembled DNA computer can be built, and more than 10 years before such a computer could be ready for practical use, said Dwyer.

Eventually self-assembly could replace photolithography as the best way to fabricate computing devices, said Dwyer. It may prove more practical, however, for the two methods to merge into a hybrid fabrication method, he said.

Dwyer's research colleagues were John Poulton of Rambus Inc., and Russell Taylor and Leandra Vicci of the University of North Carolina at Chapel Hill. The work appeared in the October 28, 2004 issue of Nanotechnology. The research was funded by the National Science Foundation (NSF).

Timeline:   > 10 years
Funding:   Government
TRN Categories:  Architecture; DNA Technologies; Nanotechnology
Story Type:   News
Related Elements:  Technical paper, "DNA Self-Assembled Parallel Computer Architectures," October 28, 2004, Nanotechnology


January 12/19, 2005

Page One

Video organizes paper

Conversations control computers

DNA scheme builds computers

The History Files:
A Short History of the Computer

Letter to readers

Copy-and-paste goes natural
RNA tiles form nanopatterns
Input device tracks muscle tremors
Nano gas turbine designed
Ultrasound makes blood stand out
Silicon surfaces speed circuits
Branchy molecules make precise pores


Research News Roundup
Research Watch blog

View from the High Ground Q&A
How It Works

RSS Feeds:
News  | Blog  | Books 

Ad links:
Buy an ad link


Ad links: Clear History

Buy an ad link

Home     Archive     Resources    Feeds     Offline Publications     Glossary
TRN Finder     Research Dir.    Events Dir.      Researchers     Bookshelf
   Contribute      Under Development     T-shirts etc.     Classifieds
Forum    Comments    Feedback     About TRN

© Copyright Technology Research News, LLC 2000-2006. All rights reserved.