Most likely sending pseudorandom data so that the data can be validated at the other end.
Given they say it’s really 19 fibers in one, that’s really just 6,600Gb/s per fiber which is really just 4 colors per fiber with one of those and some amplifiers: https://www.fs.com/c/1.6t-osfp-infiniband-1392
Apparently those go into a watercooled switch. Those 1.6T NICs sound absolutely insane. Makes your home 10G network look strings and cans.
It’s not that insane in perspective. Probably still needs a whole rack of equipment to run just that test, but the technology is not too far off that it’s quite plausible.
they say it’s really 19 fibers in one … It’s not that insane in perspective
The impressive bit being that the bundle of 19 fibres is around the same overall diameter as a single regular one - “diameter of five-thousandths of an inch (0.127 millimeters), which is the same thickness as most existing single-fiber cables already in use” - meaning those individual strands are unbelievably thin.
It’s going to be interesting to see how a cable like that is getting fixed in the field when a backhoe inevitably goes om nom nom on one.
Ignoring clickbait title, this is impressive. Networked devices used to be the limit on data transfer.
Are there any devices even capable at reading/writing at 125,000G/sec?
Seems breakthroughs here are more relevant to for backhaul networks.
Most likely sending pseudorandom data so that the data can be validated at the other end.
Given they say it’s really 19 fibers in one, that’s really just 6,600Gb/s per fiber which is really just 4 colors per fiber with one of those and some amplifiers: https://www.fs.com/c/1.6t-osfp-infiniband-1392
Apparently those go into a watercooled switch. Those 1.6T NICs sound absolutely insane. Makes your home 10G network look strings and cans.
It’s not that insane in perspective. Probably still needs a whole rack of equipment to run just that test, but the technology is not too far off that it’s quite plausible.
The impressive bit being that the bundle of 19 fibres is around the same overall diameter as a single regular one - “diameter of five-thousandths of an inch (0.127 millimeters), which is the same thickness as most existing single-fiber cables already in use” - meaning those individual strands are unbelievably thin.
It’s going to be interesting to see how a cable like that is getting fixed in the field when a backhoe inevitably goes om nom nom on one.
0.125 P/s