Quote:
Originally Posted by Pierre
And DCF, but Bragg gratings have improved.
Latency in general is just to do with how long the cable is.
A new low latency trans-Atlantic cable is currently being installed, it's secret? It's shorter than all the others (and fewer repeaters).
A few milliseconds advantage for an algorithmic Stoke trade = £100millions
|
Yeah, the number I hear being bandied about is 1ms of latency is worth about £5m a year to HFT companies.
The transatlantic paths are getting quite close to the theoretical maximum speed now based on the speed of light, with previous gen cables still giving round-trip latencies 20-40% higher than the cable length alone would suggest. The claimed 55-60ms RTT from London to NYC on the latest gen cables come within 10% of the absolute best we'll ever get, unless someone embeds an open-air a waveguide through the earth's mantle.
However paths on the other side of the world are still slow enough that less than half the round trip time is attributable to actual signal propagation along the fibre.
But I digress. In the home broadband environment, most of the latency in the last mile comes from the bridging devices at either end - i.e. the modems, DSLAMs, and CMTS.