Agreed. Actual ping would not work, if we're talking about unixy system sending ping. Active devices increase latency in single digit microsecond, the host latency is much much higher, so the ping results would largely measure your host latency variation. You'd need NIC with hardware timestamping to get anything useful. I don't think we have unixy 'ping' tool which uses that? Even though typical server NIC today does do that, unfortunately SR-IOV does not expose it to guests, so guests would need direct NIC access to utilise it. On Tue, 30 Oct 2018 at 10:03, Ben Cannon <ben@6by7.net> wrote:
I’d be astonished if they varied much from lightspeed.
-Ben
On Oct 29, 2018, at 8:11 PM, Mehmet Akcin <mehmet@akcin.net> wrote:
Hello there
I have been searching for CLS (Cable Landing Station) to CLS latency infirnation for submarine cables. I was able to identify only a handful of them via sales catalogs. I am still missing like 320~ systems, does wnyone have this info available?
If i can’t find the data, i am going to formulate from lentgh of fiber diveded by speed of light on fibre (~ +\- resistance) but I really would prefer actual ping ;)
This is for www.networkatlas.org project. Thank you in advance!
Mehmet -- Mehmet +1-424-298-1903
-- ++ytti