One gram of DNA contains about 10^21 nucleotide bases which is equivalent to
the information storage capacity of a billion terabits. In addition, the
massively parallel biochemical reaction in DNA solution has the potential of
supercomputing at the nanoscale with extremely low energy consumption. These
possibilities of DNA-based thermodynamic computing have been investigated
internationally by several research groups since Adleman's demonstration of
its feasiblity in 1994. In this talk, we present the recent advances in DNA
computing technology, its opportunities and challenges. Concrete examples
will be taken from our experimental results of the Molecular Evolutionary
Computing (MEC) project supported by the Korean Ministry of Industry and
Energy. We also discuss some research issues to be addressed by physicists
and chemists as well as computer scientists and molecular biologists.