[Pool] RIPE Atlas monitoring pool

Daniel Frank ntp at tokuko.de
Sun May 31 11:29:16 UTC 2015

Am Freitag, 29. Mai 2015, 11:02:47 schrieb Philip Homburg:
> Hi,
> I work on the RIPE Atlas project (https://atlas.ripe.net/). We have more
> than 8000 probes world wide that can performance various types of
> measurements. Ping, traceroute, dns, http, but also ntp.
> My plan is to start a measurement where all probes perform ntp 
> towards pool. Especially with the upcoming leap second, now seems to 
> an interesting time to start the measurement. There is almost a month 
> collect a baseline reference. And some system might announce the leap
> second early.
> My current plan is to have each probe run an ntp measurement toward
> pool.ntp.org every 15 minutes. A measurement involves looking up
> pool.ntp.org in DNS and sending 3 requests towards the first IPv4 
> that is returned. (IPv6 capable probes would run a separate 
> to the first IPv6 address).

I'm not really sure what you want to do with the data you gather. The result 
would be pretty much similar to what the pool monitoring system already 
gathers, just as seen from an outsider because you will get a different 
server on every measurement.
Ok you can make a statistic about the leap second e.g. to show roughly 
how the percentage of servers announcing the leap second is. But is that 
all you want to accomplish?
Personally I don't mind this, as long as you don't have all 8000 probes 
start at once but randomly distribute the load over the 15 minutes.
8000 probes multiplied with 3 packets at once is about 1.7 Megabyte and 
could easily annoy servers with weaker links.


More information about the pool mailing list