Big latency change
Hi folks, How can I check what happened at 13:00 (Dont know if the time in my probe is UTC or UTC+2) of today in the a.root-servers.net? My latency with the a-root changed from 117 to 45ms and Im curious of what happened. I've took a look in the RIPEstat routing for the IP Address of the root server, and at the same time the BGP had changes, but there are too many AS involved I cant follow it. Thanks in advance! -- Daniel Baeza
On 2015-05-15 18:28, Daniel Baeza (Red y Sistemas TVT) wrote:
Hi folks,
How can I check what happened at 13:00 (Dont know if the time in my probe is UTC or UTC+2) of today in the a.root-servers.net?
My latency with the a-root changed from 117 to 45ms and Im curious of what happened.
I've took a look in the RIPEstat routing for the IP Address of the root server, and at the same time the BGP had changes, but there are too many AS involved I cant follow it.
Thanks in advance!
Hi, a-root is anycasted, so most likely your probe switched to a different instance of the server. You can see their instances here: https://atlas.ripe.net/contrib/root_anycast.html?msm_id=9 You can check dns measurements your probe does and see if this is indeed the case -- they are available to you in the probe status page, built-in measurements tab. Perhaps "v4 DNS to a.root-servers.net (UDP hostname.bind)" is the best to try. Similarly, traceroutes could tell you exactly what changed. (We report all times in UTC) Cheers, Robert
Hi Robert, Thanks for you answer but the only built-in graphs are for ping. The v4/v6 dns are only available for download in json. Also, how can I see an "old" traceroute, I mean the one before the latency did go down? Regards, El 15/05/2015 a las 19:37, Robert Kisteleki escribió:
On 2015-05-15 18:28, Daniel Baeza (Red y Sistemas TVT) wrote:
Hi folks,
How can I check what happened at 13:00 (Dont know if the time in my probe is UTC or UTC+2) of today in the a.root-servers.net?
My latency with the a-root changed from 117 to 45ms and Im curious of what happened.
I've took a look in the RIPEstat routing for the IP Address of the root server, and at the same time the BGP had changes, but there are too many AS involved I cant follow it.
Thanks in advance!
Hi,
a-root is anycasted, so most likely your probe switched to a different instance of the server. You can see their instances here: https://atlas.ripe.net/contrib/root_anycast.html?msm_id=9
You can check dns measurements your probe does and see if this is indeed the case -- they are available to you in the probe status page, built-in measurements tab. Perhaps "v4 DNS to a.root-servers.net (UDP hostname.bind)" is the best to try. Similarly, traceroutes could tell you exactly what changed.
(We report all times in UTC)
Cheers, Robert
-- Daniel Baeza
Hola Daniel My probe (#10611) for instance also saw a change in routing towards A-root around the same time [1]. On Sat, May 16, 2015 at 8:58 AM, Daniel Baeza (Red y Sistemas TVT) <d.baeza@tvt-datos.es> wrote:
Hi Robert,
Thanks for you answer but the only built-in graphs are for ping. The v4/v6 dns are only available for download in json.
Indeed, built-in graphs only visualise ping latencies, but it is not hard to get what you want out of the raw JSON files. Please see [2]. Apparently, I was no longer hitting their site at NYC but LON.
Also, how can I see an "old" traceroute, I mean the one before the latency did go down?
The measurement API allows you to download measurement results in bulk, but also allows you to filter what data you're interested in (which probe, start and stop datetime, etc), please see [3]. If you want older traceroute data, just tell the API in which time window you are interested. HTH, [1] https://stat.ripe.net/m/widget/atlas-ping-measurements#w.mode=condensed&w.measurement_id=1009&w.probe_id=10611&w.starttime=2015-05-09T09%3A03%3A17&w.endtime=2015-05-16T09%3A03%3A17&w.resolution=1h [2] [ops@eh ~]$ curl -so- 'https://atlas.ripe.net/api/v1/measurement/10309/result/?start=1431693000&stop=1431696600&prb_id=10611&format=json' | jq .[].result.answers[].RDATA "nnn1-nyc3" "nnn1-lon3" [3] https://atlas.ripe.net/docs/rest/ -- "If you want to go fast, go alone. If you want to go far, go together."
I can see a similar change on the 3 Probes I have access to (1x V1 + 1 Anchor, close to each other, one on a different upstream). The maybe more interesting thing is that for 2 out of the 3, *after* the decrease of RTT, there were short bursts of RTT > than the previouse ~100+ : ~250ms and ~300ms. So, whatever the change was, it would be intersting to ask those folks if the change was intended or a side-effect of another event in the 'net... On 2015-05-15 18:28, Daniel Baeza (Red y Sistemas TVT) wrote:
Hi folks,
How can I check what happened at 13:00 (Dont know if the time in my probe is UTC or UTC+2) of today in the a.root-servers.net?
My latency with the a-root changed from 117 to 45ms and Im curious of what happened.
I've took a look in the RIPEstat routing for the IP Address of the root server, and at the same time the BGP had changes, but there are too many AS involved I cant follow it.
Thanks in advance!
FWIW, Wilfried
participants (4)
-
Daniel Baeza (Red y Sistemas TVT) -
Iñigo Ortiz de Urbina -
Robert Kisteleki -
Wilfried Woeber