Re: beyond _tera_[Fwd: dozenal and hexadecimal digits]

> Date:         Mon, 15 May 2000 08:34:32 -0400
> From: John Cowan <jcowan@...>
> I once read a brief squib (in Datamation, perhaps?) about the possibilities
> of yottabit/sec networking.  It would require fiber-optic cables about 12 m
> (almost 40 feet) in diameter, which could be laid along disused railway tracks.
OK, silly math time...

Let's say we can build waveguides of buckytubes, at 100Å diameter.
These will transmit signals at about the same wavelength, i.e. 30 PHz.
One square millimetre of cable will hold 117 million tubes for an
aggregate bandwidth of 3.53 YHz; in other words, if we can modulate at
20 PHz we can send 1 Yib/s (yobibyte per second) through a cable 2/3
of a millimetre thick.

Assuming that we only need one photon per bit, at that frequency
(borderline X-rays) we get 6.3e-17 J/bit or 7.6 MJ/Yib. Keeping the
cable from melting, or the C-C bonds from ionizing away, will be a
nice engineering challenge.

If you want visible light (400 nm/750 THz) the cable would need to be
17 cm thick. But you'd get by with 190 kW.

Lars Mathiesen (U of Copenhagen CS Dep) <thorinn@...> (Humour NOT marked)