Quote from GreenMan:
By âfastestâ I mean the the lowest delay between tick being generated at the exchange till it come to my PC (in Europe). Especially at the moments of releasing the economic data.
Quote from PocketChange:
Colocate a server at Equinix Chicago and process the data there for sub 3ms. Transporting the tick data to Europe adds approx 100ms latency. Processing time sensitive orders requires close proximity to your broker's infrastructure.
Quote from GreenMan:
There is no reason to colocate in US because I'm autotrading at european exchanges. So I'm not concerned about 3-5 ms it takes for data to travel in US. What I really need is to be first to receive S&P 500 futures data in europe. How can this problem be solved?
Quote from GreenMan:
There is no reason to colocate in US because I'm autotrading at european exchanges. So I'm not concerned about 3-5 ms it takes for data to travel in US. What I really need is to be first to receive S&P 500 futures data in europe. How can this problem be solved?
Quote from Jerkstore:
The longer the race, the wider the time gap between the fastest networks, and everyone else.
If you want to be the fastest you have to write to an exchange directly. In your case, you want to write to the CME Globex API. Collect the data in a server at the Chicago Equinex center (or talk to the CME about a pure co-location). Call a company like IPC to lease a line from Chicago to the exchange server center in Europe.
This require a lot of coding, networking expertise, as well as high monthly payments for lines. Implement a two-tier approach to minimize costs. When developing your system be sure that you use the super fast/expensive lines for only the most time-sensitive data. Use your regular data provider for everything else.