placing a probe into wireless networks (not terrestrial)
Dear Atlas-ml-followers, this a summary of a brief discussion I had with Michel Stam during RIPE85. We believe there might be a great value putting up a public Internet monitoring probe (vantage point) into low earth orbit satellite systems, such as Starlink. Monitoring from terrestrial end-points delivers high variances, due to frequent sat-changes. If we could put up an ATLAS-probe (be it SW - which is probably easier than HW) into a specific plane, we could add interesting tests and try to interpret the values depending on position over specific areas, atmospheric influences, solar influences as well as other effects that would change signal patterns. I would be offering support to try to establish such a probe and help to setup a project defition. Any thoughts or feedback to this idea? Best regards Kurt Kayser
Hi Kurt, Thanks for a post, it was good having a chat. What I would like to know, on a more high-level note, is how the community sees measuring the Internet via wireless channels. This brings some challenges, because measurements can be affected by such as interference (weather, radio, solar, etc) and also the probe itself moving. That of course does not detract from the initial question of deploying probes on LEO networks, but rather should be seen as adding on to it. Regards, Michel
On 28 Oct 2022, at 10:04, Kurt Kayser <Kurt.Kayser@online.de> wrote:
Dear Atlas-ml-followers,
this a summary of a brief discussion I had with Michel Stam during RIPE85.
We believe there might be a great value putting up a public Internet monitoring probe (vantage point) into low earth orbit satellite systems, such as Starlink.
Monitoring from terrestrial end-points delivers high variances, due to frequent sat-changes.
If we could put up an ATLAS-probe (be it SW - which is probably easier than HW) into a specific plane, we could add interesting tests and try to interpret the values depending on position over specific areas, atmospheric influences, solar influences as well as other effects that would change signal patterns.
I would be offering support to try to establish such a probe and help to setup a project defition.
Any thoughts or feedback to this idea?
Best regards
Kurt Kayser
-- ripe-atlas mailing list ripe-atlas@ripe.net https://lists.ripe.net/mailman/listinfo/ripe-atlas
On 2022-10-28 10:04, Kurt Kayser wrote:
I would be offering support to try to establish such a probe and help to setup a project defition.
Any thoughts or feedback to this idea?
In addition to what's already been said: it's useful for the whole community to increase the diversity of the network. So, the more networks we can have probes in, the merrier! While doing so I find it highly useful to add proper meta-data as well (ie. probe tags) to aid people understanding what they observe. For example it is confusing to see high RTTs for a probe, unless it's documented that it's on a (non-LEO) satellite link for example. Side-note: at the moment we have 31 probes connected on SPACEX-STARLINK (AS14593), see https://atlas.ripe.net/results/maps/network-coverage/?filter=AS14593 Regards, Robert
Hi All, Giving the topic a bit of a poke again; After some internal discussion, wireless probes in general could be tagged by looking at particular AS numbers (this excludes wifi of course). Specifically to StarLink, this could be possible but would require further understanding of the way internet measurement would work with these satellites. Does anyone have insights that they could share in this regard? Regards, Michel
On 28 Oct 2022, at 10:32, Robert Kisteleki <robert@ripe.net> wrote:
On 2022-10-28 10:04, Kurt Kayser wrote:
I would be offering support to try to establish such a probe and help to setup a project defition. Any thoughts or feedback to this idea?
In addition to what's already been said: it's useful for the whole community to increase the diversity of the network. So, the more networks we can have probes in, the merrier!
While doing so I find it highly useful to add proper meta-data as well (ie. probe tags) to aid people understanding what they observe. For example it is confusing to see high RTTs for a probe, unless it's documented that it's on a (non-LEO) satellite link for example.
Side-note: at the moment we have 31 probes connected on SPACEX-STARLINK (AS14593), see https://atlas.ripe.net/results/maps/network-coverage/?filter=AS14593
Regards, Robert
-- ripe-atlas mailing list ripe-atlas@ripe.net https://lists.ripe.net/mailman/listinfo/ripe-atlas
participants (3)
-
Kurt Kayser -
Michel Stam -
Robert Kisteleki