Watched a movie on a flight called The Hummingbird Project. It’s called that because one flap of a hummingbird’s wings takes 16ms.
That is the time these quants at a HFT firm targets for a private fiber optic network from KS to NY, which would be the shortest line possible. Well it includes code execution time to run this front running software to process thousands of HFT transactions per second.
So these guys quit the firm, find backers and they deploy dozens of crews to trench a direct line. That means paying off dozens of property owners to trench a 12-inch wide conduit. Where they get into difficulty is in Appalachia, where they have trouble drilling through solid granite. Oh there’s a national park there but somehow, their investors are connected enough to allow all those crews with big equipment to drill in the park.
Anyways, their former employer hires some kid out of NYU to solve an obstacle for direct microwave link over the same terrain and they do it in 11 ms.
Are a series of microwave towers faster than fiber optic? Wouldn’t rain and other weather degrade performance?
Of course this is Hollywood so accuracy wouldn’t be expected.
Other part is this software, where the genius who came up with this scheme has to shave 1.39 ms in code execution, so the movie shows him poring over code. Maybe throw faster hardware at it? Or parallelize the code and throw clusters of compute power?
In the end they’re defeated by this microwave network but they hit their target.
So the genius guy comes up with another idea. A “neutrino messaging system” where they’d make an accelerator which goes in a straight line under the earth crust, avoid added distance from curvature of earth, and they’d shoot neutrinos.
That is the time these quants at a HFT firm targets for a private fiber optic network from KS to NY, which would be the shortest line possible. Well it includes code execution time to run this front running software to process thousands of HFT transactions per second.
So these guys quit the firm, find backers and they deploy dozens of crews to trench a direct line. That means paying off dozens of property owners to trench a 12-inch wide conduit. Where they get into difficulty is in Appalachia, where they have trouble drilling through solid granite. Oh there’s a national park there but somehow, their investors are connected enough to allow all those crews with big equipment to drill in the park.
Anyways, their former employer hires some kid out of NYU to solve an obstacle for direct microwave link over the same terrain and they do it in 11 ms.
Are a series of microwave towers faster than fiber optic? Wouldn’t rain and other weather degrade performance?
Of course this is Hollywood so accuracy wouldn’t be expected.
Other part is this software, where the genius who came up with this scheme has to shave 1.39 ms in code execution, so the movie shows him poring over code. Maybe throw faster hardware at it? Or parallelize the code and throw clusters of compute power?
In the end they’re defeated by this microwave network but they hit their target.
So the genius guy comes up with another idea. A “neutrino messaging system” where they’d make an accelerator which goes in a straight line under the earth crust, avoid added distance from curvature of earth, and they’d shoot neutrinos.
