3 ms·

It would be an interesting problem to really try and get the timestamps around 10's of ms for the time sorting bound. With multiple datacenters you might get up

by timf 16y ago

It would be an interesting problem to really try and get the timestamps around 10's of ms for the time sorting bound. With multiple datacenters you might get up to ~10ms drift with NTPv4 so you're starting point is already pretty crippled. And how would you even test? :-)

kingryan 16y ago
NTP seems to work fairly well for this. It takes into account network latencies and attenuates over time.
- timf 16y ago
  Well, quoting from: http://www.eecis.udel.edu/~mills/ntp.html http://www.eecis.udel.edu/~mills/ntp.html "Used in the Internet of today with computers ranging from personal workstations to supercomputers, NTP provides accuracies generally in the range of a millisecond in LANs and up to a few tens of milliseconds in the global Internet" But this is news to me, interesting: "When kernel support for precision timing signals, such as a pulse-per-second (PPS) signal, is available the accuracy can be improved ultimately to the order of one nanosecond in time and one nanosecond per second in frequency."
profquail 16y ago
You could use a GPS clock in each datacenter...AFAIK, the time signal is accurate to ~1 µs, so you shouldn't need to coordinate the time signal between datacenters.
adulau 16y ago
Yes, this is a good point. There is a pretty alternative to NTP and limiting the drift called radclock : http://www.cubinlab.ee.unimelb.edu.au/radclock/ http://www.cubinlab.ee.unimelb.edu.au/radclock/ more information about the accuracy is available there : http://www.cubinlab.ee.unimelb.edu.au/radclock/performance.php http://www.cubinlab.ee.unimelb.edu.au/radclock/performance.p... When looking at it, I was a bit dubious until we made some tests at work and compared it with our current NTP using GPS IRIG-B receiver. Even for disconnect period longer than 48 hours, we had a small drift (around 200 nanoseconds) compared to traditional NTP. The only drawback is that you need a kernel path to make it works.

by timf 16y ago

kingryan 16y ago

NTP seems to work fairly well for this. It takes into account network latencies and attenuates over time.

timf 16y ago
Well, quoting from: http://www.eecis.udel.edu/~mills/ntp.html http://www.eecis.udel.edu/~mills/ntp.html "Used in the Internet of today with computers ranging from personal workstations to supercomputers, NTP provides accuracies generally in the range of a millisecond in LANs and up to a few tens of milliseconds in the global Internet" But this is news to me, interesting: "When kernel support for precision timing signals, such as a pulse-per-second (PPS) signal, is available the accuracy can be improved ultimately to the order of one nanosecond in time and one nanosecond per second in frequency."

profquail 16y ago

You could use a GPS clock in each datacenter...AFAIK, the time signal is accurate to ~1 µs, so you shouldn't need to coordinate the time signal between datacenters.

adulau 16y ago

Yes, this is a good point. There is a pretty alternative to NTP and limiting the drift called radclock : http://www.cubinlab.ee.unimelb.edu.au/radclock/ http://www.cubinlab.ee.unimelb.edu.au/radclock/ more information about the accuracy is available there : http://www.cubinlab.ee.unimelb.edu.au/radclock/performance.php http://www.cubinlab.ee.unimelb.edu.au/radclock/performance.p... When looking at it, I was a bit dubious until we made some tests at work and compared it with our current NTP using GPS IRIG-B receiver. Even for disconnect period longer than 48 hours, we had a small drift (around 200 nanoseconds) compared to traditional NTP. The only drawback is that you need a kernel path to make it works.