>>13877421 >>13877416 >>13877410 >>13877365 >>13877344Telecom guy here who owns a FTTH isp.
The fastest single frequency encoding commonly used in fiber optics today is 56Gbit/s, we however use muxers and different frequency lasers to stack multiples of these on a single fiber called WDM Wavelength-division multiplexing. With commodity hardware we can get 80 channels between 1514nm and 1577nm, with slightly more exotic stuff people get 160 channels in the same range. About 4.5 tbps on a single fiber with commodity shit I can buy today.
However single mode fiber is good from about 1260nm to 1700nm, some of these frequencies in this range need special glass to work with because water that gets absorbed in the glass and has absorption issues causing loss but they sell glass that is formulated to not have that issue.
And all this is essentially just bit banging the traffic through, on and off really fast some of the new 112Gbit stuff that will be commodity grade some time next year is using PAM4 where they send four different intensities of light to get four bits of data per symbol on one laser frequency.
Many of the issues we experience with backscatter (return loss) chromatic dispersion etc would not exist in free space through a vacuum.
The absolute brightest optics we use in telecom is 3mw of optical power and even that is quite rare and can easily reach 120km, when needed we use erbium-doped fiber amplifiers to boost the signal, this same technology is good up to tens of Kw.
I don't think it will be hard to get multiple terabits of traffic between earth and mars orbits even for the earliest colonies. The needed intensity levels on the lasers are there, the needed sensitivity of the receivers are there. These things will be far easier to aim than orbital space telescopes.
With our current technology using 25ghz channel spacing from 30thz to 330thz gives you about 12,000 channels, with a symbol rate of 53ghz and PAM4 that would be about 2,500 Tibps
