Using RIPE Atlas for NTP Measurements

If you are operating a public available NTP server, for example when you’re going to join the NTP Pool Project, you probably want to test whether your server is working correctly. Either with a one-off measurement from hundreds of clients or continuously to keep track of its performance. You can use the RIPE Atlas measurement platform (Wikipedia) for both use cases. Here’s how:

This article is one of many blogposts within this NTP series. Please have a look!

I am hosting two RIPE Atlas probes in different ASes for many years. (By the way: They are the only probes in both ASes.) Hence I have enough credits to run several ongoing tests as well. Have a look at my blogpost about an overview of RIPE Atlas measurements and some stats.

Create a New Measurement

The first step is to create a new measurement. In this example I am testing my Meinberg LANTIME M200 appliance “”, which is IPv6-only, hence the address family selection of IPv6. To have a realistic test from all over the world, I am checking the “resolve on probe” box. Note that the “interval” section will disappear when I am enabling the one-off measurement in the third step below:

In the second step, I am selecting the probes. I deleted the default “Worldwide 10” selection and added a manual set with 500 probes worldwide while including the “system-ipv6-works” tag. Otherwise, you’ll have too many errors from not working IPv6 probes:

Finally, I am using a one-off timing and creating the measurement:


Since the measurements are publicly available you can have a look at this just created measurement #20366117 by yourself. Initially, it takes a couple of minutes (of course) until the results are available. In the end, the results were as expected: Very low offsets from all over the world, while a few probes did not succeed at all:

Ongoing Tests & other Examples

Besides this one-off measurements, I have a couple of ongoing tests as well, for example, measurement #22810167 which uses 50 probes worldwide to query my NTP server every 900 seconds. Also note that with the help of these RIPE Atlas measurements I encountered an initial problem with my load balancing setup for NTP, refer to Load Balancing NTP via F5 BIG-IP LTM.

Featured image “oranges – en masse” by Georgie Sharp is licensed under CC BY-NC 2.0.

Leave a Reply

Your email address will not be published. Required fields are marked *