Best DC for AU? Tokyo?
I think some ISP's in AU may route data via the USA to tokyo, does anyone know which ones would do this? and what the speed impact is?
5 Replies
@webconcepts:
I am getting 125ms pings to it, how long has this been setup for now?
Officially, since Monday afternoon. However, it has probably been in testing for some time now.
> I think some ISP's in AU may route data via the USA to tokyo, does anyone know which ones would do this? and what the speed impact is?
There's another forum post where someone reported their Australian ISP routed traffic via California. Assuming the route is symmetrical (i.e. traffic to that ISP from Tokyo goes via California), a round-trip will involve four trips across the Pacific Ocean. This usually doesn't help performance, especially relative to Fremont (which is within light-nanoseconds of where the trans-oceanic cables tend to land).
On the other hand, others have found more sensible paths to Tokyo.
Your mileage (or timing) may vary, but your speed of light will not.
Most ISP's have agreements with the main cables out of Australia so in theory you should always get the best route out of Australia.
In My case, my provider only has a single link across the southern cross cable, which means the route cannot use the AJC which is the best link from Australia (via Guam) to Japan.
@hoopycat:
Your mileage (or timing) may vary, but your speed of light will not.
ahemahem Scientists may have found neutrinos faster than the speed of light
I want my trans-Pacific neutrino cable now.
Can we scale this up to OC-48?
@hoopycat:
60 ns per ~750 km isn't the exciting part as far as I'm concerned. Sending "information" through 750 km of rock? That's the useful part.
Can we scale this up to OC-48?
I don't know, even 60ns per 750km is enough to blow the theoretical doors wide open, and really messes around with causality.
I never did get all the mathy "FTL means effect can precede cause" stuff (nobody has ever managed to explain it in laymans terms, it seems, because giving an example of "you receive the reply before you send the message" doesn't explain why that's the case), but even the tiniest amount over the speed of light opens the door to that.
But that will almost certainly turn out to be due to flawed measurement. Neutrino communication would be neat, but would require improvements in the ability to generate neutrinos of something like seven magnitudes, and current detectors are incredibly massive; sending data to the other side of the planet isn't very useful if your detector's weight is measured in kilotons and must be buried kilometers underground just to send a hundred bits per second. Maybe some day we'll be able to improve on that enough to make it practical, but as it stands, that sounds like it'd be decades or centuries away.