> The hosts i am looking at are in AWS us-east-1
Google says that's in Northern Virginia so it isn't very far from NIST. (But
routers can do strange things.)
The question for the curious is which direction is taking the long route and
how far out of the way is it going. There is enough info in the ntpq peers
printout. The delay column is the round trip time. The offset is 1/2 the
difference in routing delays, but I can never remember the sign.
If you are collecting rawstats, the data is in there. There are 4 time
stamps. Normally, ntpd assumes the routing delays are symmetric and computes
the time offset. If you assume the clocks are correct you can compute the
routing delays. If you label the time stamps T0, T1, T2, and T3, then T1-T0
is the time from client to server and T3-T2 is the time from server back to
the client.
Again, that assumes your local clock is correct.
If you like gnuplot, try this:
grep <IP Address> rawstats-file > raw.ipaddress
plot \
"raw.ipaddress" \
using ($2/3600):(($6-$5)*1000) \
title "Out" with points lt 1 pt 5, \
"raw.ipaddress" \
using ($2/3600):(($8-$7)*1000) \
title "Back" with points lt 2 pt 5
You can try traceroute. It works better if you can get one from the other
end too. You might be able to guess where things are going from host names.
--
These are my opinions. I hate spam.
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool
