Viktor Dukhovni <[email protected]> writes: >This does not look to me like a compelling rationale. A high-frequency >trading system that does not route trades over a connection that was >established well before market open, and expects to beat the competition by >minimising connection establishment latency, is perhaps doing it wrong. For >already established TLS connections latency does not depend on which key >agreement group was used in the initial handshake.
The following is some years out of date so things may have changed since then but HFT uses (used) UDP for reporting market data because you can't have any latency and TCP for order placement because you need reliability, but even that's modded TCP with anything that would introduce delays or stalls removed. The interesting stuff isn't the software but the custom hardware used to minimise any kind of delay, at the time eyewateringly expensive FPGAs (Virtex UltraScales) but now presumably ASICs. >One might also point out that the payload of high-frequency trades is not >likely to be a long-term secret. Since HFT is sort of front-running the market but technically it's not so it's not illegal, the secret may only be valuable for a few hundreds of milliseconds and even if you discover it you can't react fast enough to act on it yourself. As you mention above, the logical thing to do would be to set up the TLS connection before market open and not hope that your TLS handshake completes in time to get your urgent buy order through. Peter. _______________________________________________ TLS mailing list -- [email protected] To unsubscribe send an email to [email protected]
