# Microsecond trading with neutrinos

The Spread Networks corporation recently laid down 825 miles of fiberoptic cable between New York and Chicago, stretching across Pennsylvania, for the sole purpose of reducing the latency of microsecond trades to less than 13.33 milliseconds (http://www.spreadnetworks.com/spread-networks/spread-solutions/dark-fiber-networks/overview). The lesson I would draw from this is that, in the near future, oil and natural gas extraction won't be the only lucrative use of ocean platforms.

So here's my question - since trades are occurring on the scale of tens to hundreds of microseconds, and considering the amount of money involved, can one use neutrino beams to beat the limitation due to having to travel the great-circle/orthodromic distance between two trading hubs? I'm imagining something similar to the MINOS detector (http://en.wikipedia.org/wiki/MINOS), where a neutron beam was generated at Fermilab in Batavia, Illinois, and detected ~735 km away, ~700 meters under the ground in a Northern Minnesota mine.

Is it possible to beat a signal traveling at the speed of light across the great-circle distance from, say, New York to Tokyo, using a neutron beam traveling the earth? Is it realistic to talk about generating these beams on a microsecond time-scale?

Addendum - Over what distances can you reasonably detect a neutrino beam?

-
Considering how hard it is to detect neutrinos, the advantages of having a shorter distance would be largely offset by the size of the detectors you'd need to read out messages. I don't think we even have enough control on neutrinos to even form a message to start with. I think your post belongs in scifi.SE . –  Raskolnikov Jul 27 '11 at 8:04
@ Raskolnikov - neutrinos can be generated on demand by muon decay. –  Richard Terrett Jul 27 '11 at 8:13
@Richard Oh yes and it is possible to modulate that beam up to GHz frequencies with a shutter, right? –  Georg Jul 27 '11 at 8:44
""for the sole purpose of reducing the latency of microsecond trades to less than 13.33 milliseconds"" Mhmmm microsecond trades down to 13 milliseconds? Strange. –  Georg Jul 27 '11 at 8:47
@user8861""for the sole purpose of reducing the latency of microsecond trades to less than 13.33 milliseconds "" Meanwhile I looked up that link, I could not find that the cable was laid for that "sole purpose" –  Georg Jul 27 '11 at 8:56

## 1 Answer

Whether or not neutrinos would be suitable for rapid trading, people have seriously considered their utility for signalling in difficult environments. I read an article a while back about a paper (published in Phys. Lett. B, but I can't access that from here) by Patrick Huber which proposed using neutrinos for through-the-earth communication to submarines as an alternative to ELF, where bandwidths become competitive. The submarine would pick up the modulated cherenkov radiation produced by the generation of muons in seawater. This certainly allows faster-than-great-circle transmission times, but this is not the reason why the technique is attractive. The preprint indicates that calculated antipode to antipode bandwidth is only 10 b/s which doesn't seem ready for high-intensity trading.

Addendum:

If we consider a continuous lossless fibre optic link between antipodes around the equator, the transmission time will be about 99 ms, whilst the through-earth travel time (at $\approx{c}$) is 42 ms. Obviously this counts for nothing if you have high-latency equipment at either end.

Whilst the improvement in transmission time hardly seems worth it, it occurs to me that this would be a useful technique for communicating between either side of a huge, highly oblate structure such as a wide but thin disk-shaped megastructure, however that's veering in to sci-fi territory.

-
With the modern state of the art I doubt you could extract a high-bandwidth signal from a neutrino detector in tens of milli-seconds. With a ~13,000 km distance you are going to need either a huge detector or a scorching hot beam to get more than a few neutrinos per second. (There is a very nice experiment that could be run with a ~10,000 km baseline, if only we could get the rate; effort is being put into a shorter but harder baseline because the rates are achievable). Make the detector IceCube sized and the readout latencies are going to kill you. –  dmckee Jul 27 '11 at 23:31
@ dmckee - That's correct. As mentioned, Huber's calculations give an antipode to antipode bandwidth of only 10 bits per second, where as ELF would only give a bit per minute or so. At the receiving end, you would essentially be using the combination of ocean and sensor-studded submarine as a kamiokande-style detector. The tightly collimated beam works very much to the advantage of the system here. –  Richard Terrett Jul 28 '11 at 4:38