Can free space communications cut it in the supercomputing arena?
Free space communication is a concept which has been around for many years; the ancient Greeks used their shields to flash messages amongst each other during battles. And Alexander Graham Bell developed the technology to send sound on a beam of light, something which he described as his 'greatest achievement'.
By its very nature, free space communication covers a wide range of applications – at the short range end of things, there are IrDA links and Near Field Communications. At the other end of the scale, there have been ideas about using lasers to ping information between satellites.
But the technique might now find application in supercomputing. Shunting data to the various parts of a supercomputer is a complicated business; what if you could send the information to the exact point it's needed? That's what an international team is now researching as part of a larger European high performance computing project .
Can they develop a system in which steerable rf antennas link the boards in a system? And if they can, what might that mean for supercomputing performance?
