All - I wonder whether this is a good idea or rather a really bad one... I have a free port on a Freifunk router where I could plug in a RIPE Atlas probe. Freifunk is a community where you - amongst other things - can offer a bit of your bandwidth to "the world". This connectivity sharing however is set up as a tunnel to various ISPs (for-profit companies and non-profit associations) as these would enjoy the "provider privilege": that in short is not to be held accountable for e.g. copy right infringements etc. ("Störerhaftung"). Anyhow: as part of this tunnel setup, the other end of the tunnel changes every 24 to 48 hours - and so does the upstream and therefore also the AS of the upstream provider. And now I am asking myself: a probe with ever changing upstreams/ASs - shall such an addition to the RIPE Atlas network be considered a fault or a feature? Any comments, anybody? Cheers, -C.
Hi Carsten,
Anyhow: as part of this tunnel setup, the other end of the tunnel changes every 24 to 48 hours - and so does the upstream and therefore also the AS of the upstream provider.
I think the short answer is: probably not highly useful. But of course it depends :-) Each result is tagged with the (then) source IP of the probe, as far as the infrastructure knows that. For argument's sake let's say that is the IP where the probe is otherwise connecting from. If that IP, because of the tunnels, changes regularly, then it's ... kind of ok, results from different days will vary both in RTT and in source IP (and in actual content. eg. DNS or TLS results). However, if the probe's IP stays stable but the results vary each day, then it's may be confusing to interpret. At the end of the day, measurements involving this probe would give varying results depending on time. It's totally real, but is it useful? Cheers, Robert
Hi Robert, hmm... now what? Shall we give it a try? What I could do for start is to relocate my probe #159 to this Freifunk port, say for a week or so - and you guys would keep a close eye on it wrt. the results it delivers during this period. What do you think? Cheers, -C. On 03.08.2023 19:53, Robert Kisteleki wrote:
Hi Carsten,
Anyhow: as part of this tunnel setup, the other end of the tunnel changes every 24 to 48 hours - and so does the upstream and therefore also the AS of the upstream provider.
I think the short answer is: probably not highly useful. But of course it depends :-)
Each result is tagged with the (then) source IP of the probe, as far as the infrastructure knows that. For argument's sake let's say that is the IP where the probe is otherwise connecting from.
If that IP, because of the tunnels, changes regularly, then it's ... kind of ok, results from different days will vary both in RTT and in source IP (and in actual content. eg. DNS or TLS results). However, if the probe's IP stays stable but the results vary each day, then it's may be confusing to interpret.
At the end of the day, measurements involving this probe would give varying results depending on time. It's totally real, but is it useful?
Cheers, Robert
On 3 Aug 2023, at 18:53, Robert Kisteleki wrote:
At the end of the day, measurements involving this probe would give varying results depending on time. It's totally real, but is it useful?
Hmm. If Carsten (or anyone else who may wish to offer a bit of [] bandwidth to "the world") would find it useful to know something about the availability and performance of their Freifunk "contribution", then that would answer the question, at least from one POV. OTOH, people running measurements may find it frustrates their intent. Would a new system tag, automagically added when systematic AS-hopping is observed, make sense? Dunno really. €0,02 Niall PS. "Ceci, c'est peut-être du n'importe-quoi" is so much more elegant than "Maybe this is just b-ll-x." 8-) /N
participants (3)
-
Carsten Schiefner -
Niall O'Reilly -
Robert Kisteleki