20 Jan
2016
20 Jan
'16
6:09 p.m.
On 2016-01-20 18:03, Robert Armstrong wrote:
You are absolutely right. I use physical distance as a proxy for this.
If I know better what that the link is really bad, the cost should be
higher.
However, physical distance is also not totally irrelevant. It is
"better" to talk with the US east coast than the US west coast, if I
assume that everything is pretty much equal.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
And the actual
cost is relative to the physical distance between Mim and that location.
But with links tunneled over the modern Internet, physical distance has little to do
with either bandwidth or latency. It's easily possible that I could have a slower
link talking to the guy next door than I would talking to MIM.
Seems like it would be better to set circuit costs based on either latency or
bandwidth, not distance.