satellite delay (was: Re: [or-talk] Re: Win32.Trojan.Agent appear when close Torpark)

[Sam Creasey <sammy@xxxxxxxxx>]

On Fri, Nov 17, 2006 at 06:00:43AM -0800, Total Privacy wrote:

> Another thing, that may seems off topic, but isn´t because the 
> structure of Tor system. I don´t understand why internet connections 
> around the globe is possible at all, in particular with Tor that go 
> between extra middle hops around. On the television we see reporters 
> from far away places, who´s speak are delayed one or more seconds in 
> the time of light and electricity going long ways to and from they 
> phones by some link towers, then out to space and between satellites 
> and back to eart for more link hops. 

So, you're pretty much right about the satellite thing.  A satellite
in geosynchronous orbit is about 35000km up, so a round trip to the
bird, back to earth, and then back up and down again to the original
source is at least 140,000km, and that's if your going straight up and
down (which you would normally be taking a longer path hitting the
satellite at an angle), and with 300,000km/s being the rough speed of
light, you've got at least a 1s delay after figuring in human reaction
times, etc....

> Of what reason don´t the same apply on the internet? One packet sent, 
> then wait a couple of seconds for it to reach desitination and the 
> answer packet returns, then next packet and so on. In my imagination, 
> only loading a non graphic website, or sending this email to the list 
> would take for hours, or what? But apparently it don´t... 

Ever used telnet/ssh over a satellite link?  It feels just about like
the delay you see on satellite news coverage.  Fortunately, it's
pretty rare these days.  Of course, satellites can move a ton of data,
the latency is simply quite high...

Fortunately, land based lines don't have nearly this far to go.  The
circumference of the earth at the equator is about 40000km, and it's
rare your signal would actually have to circle the entire globe.  So,
say 20000km for 1/2way around, and your speed of light delay is down
to a much more comfortable 10's of milliseconds.  (yes, light is
slower in a fiber than in a vacuum, but, still 1/10 second wouldn't
be too bad).

And the critical part of this is that it all only affects latency.
The total number of bits being pushed isn't inherently restricted by
distance (though in reality, signal attenuation, etc does play a part
in limiting total bandwidth), which is why actually moving bulk data
can still be fairly zippy.

Hope my little "Signal Propegation 101" there was at all helpful.   I
also hope it's actually correct.  :)

-- Sam