Even if I saw HFT as a valuable thing (I don't), it's unfair to create a new exemption for these people just because they have money. And if you suggest that he government shouldn't be involved, then keep in mind the only practical force stopping someone from jamming these transmissions is the exact same government bandwidth allocation guidelines.
This should not be understood as an infight by HFTs. It should be understood as the radio waves of the world being a public, shared resource and folks asking, "Are we changing these rules now?" If so, it should be done equitably.
The problem is it's not about you or me tolerating the special exception or not. It's about recourse, and we have none.
> a new exemption for these people just because they have money
The people with money can donate to all the politicians and make them aware of their needs, in exchange for the donations continuing. That's their recourse.
> This should not be understood as an infight by HFTs.
But that's exactly what it is. Multiple entities with money lobbying the same government for favorable results.
Keep in mind that the majority of back-end clearing and settlement of trades in Europe is done by Euroclear [0], who reported over 27 Trillion Euros of assets under management in 2015. In the US, virtually all back-end clearing and settlement is handled by DTCC [1], who reported handling over 1.7 Quadrillion dollars worth of value in 2011.
That’s the scale of the market you’re playing in when you first pretend to arbitrage.
[0]: https://en.wikipedia.org/wiki/Euroclear [1]: https://en.wikipedia.org/wiki/Depository_Trust_%26_Clearing_...
Latency arbitrage is old. Paul Julius Reuter (as in, Reuters) made his first fortune by using carrier pigeons to transmit trade data from Brussels to Aachen. Then he started a new agency and built telegraph lines.
Rice traders in Japan used boats/smoke-signals/runners to transmit information between rice markets in 1700s
https://digital.sandiego.edu/cgi/viewcontent.cgi?article=111... pages 15-22 has more examples.
>the shortest path
That's effectively the opposite of shortest path. If you really want the best performance you'd set up a line of microwave links.. I'm not sure how much latency the hops would add but if you have fast enough gear, it's probably not enough to justify going all the way up and back down.
Good luck setting up a microwave link between New York and London that will be faster than a short wave transmission.
i've prototyped that. :)
If you want to reach a receiver a hundred miles away an HF bounce off the E layer would get there in 0.659ms (if my math is right). If you had a 100 mile long fiber optic cable your signal would get there in 0.813ms.
Microwave repeaters can be very fast so long as you've got good positioning. If you managed a straight line of repeaters to a receiver you might be able to beat the HF path. For the best results microwave repeaters shouldn't be too far apart so for the hundred mile trip you'd want several repeaters. Each is going to add something like 150ns latency. This is cumulative with the 0.97c air latency. And it's only to a single receiver.
So microwave repeaters would beat HF sometimes, depending on topography, to a single receiver. I think HF would end up winning out more often than not and wouldn't require the infrastructure of the microwave repeaters. It would make siting receivers much easier and would likely allow receiving stations to optimize siting for the last mile to the trading systems.
That being said I don't think HFTs should be able to abuse HF by broadcasting things not usable by the general public. If they want to cut their latency they can use some of their hojillions of dollars to build microwave infrastructure or pay everyone for use of the HF airwaves they want to make money with.
The race for speed in financial markets has reached a point where no existing fiber optic or satelite based system can compete; the state of the art are short wave radio that bounce back on the ionosphere and enable intercontinental communication at the speed of light over the shortest path and with no intermediary hops. Fiber optics, for contrast, can only do about 2/3 of the speed of light and impose non-direct routes and multiple hops.