[mat-wg] Unauthenticated user access to RIPE Atlas

Sure, I am aware of RIS being used, and Atlas being used for
academia/commercial use.

But I think what I'm asking about is different, The thing I'm asking
here is that bgp.he.net is providing unauthenticated access to RIPE
Atlas probes (something the normally requires a RIPE Atlas account and
credits) that anyone on the site can request traceroute/ping/etc from.

When I asked at RIPE in Berlin I (context, I was asking with intent to
add a feature basically the same as this) was told that this was not
on, and so the bgp.tools requires users to pair up a RIPE Atlas API
key with their account before they can use the site to make RIPE Atlas
measurements, so that they are not "hiding" behind bgp.tools to use
your wording.

What I'm interested in knowing here is: Do I need to keep asking users
to attach RIPE Atlas API keys to use bgp.tools as a RIPE Atlas front
end, if HE is going to just provide open access to RIPE Atlas then
surely I don't either (and maybe there is a wider argument for not
needing logins to make RIPE atlas measurements at all)


On Tue, Oct 10, 2023 at 1:07 PM Robert Kisteleki <robert at ripe.net> wrote:
>
> Hi,
>
> I cannot speak for HE but both RIPE Atlas and RIS have a commercial use
> policy. What you want to do may fall firmly in that. IANAL but
> "commercial use" may not be a black or white thing though...
>
> More broadly speaking, we are aware that both services (and perhaps
> more) have what we can call "aggregators" as their users - entities who
> have clients/users/customers whose questions are answered (partly) by
> consulting Atlas or RIS or more. This could be for commercial or
> non-commercial reasons. In these cases on the service end we don't
> really know who the real users are and why are their questions asked.
> Having a lot of users behind a single service doesn't necessarily help
> here as this, at a minimum, skews our statistics and understanding who
> and why is using the service - because the real users can "hide" behind
> an aggregator.
>
> Perhaps there's a dialog to be had about how to improve here. With
> RIPE87 rapidly approaching, there may even be a face to face venue for
> this soon! :-)
>
> Regards,
> Robert
>
>
> On 2023-10-10 11:58, Ben Cartwright-Cox via mat-wg wrote:
> > Hello everybody!
> >
> > I was wondering around what are the rules and views on “open user
> > access” to RIPE Atlas ( as in non-authenticated users being able to
> > make RIPE atlas measurements)
> >
> > I ask because as of this morning I noticed that hurricane electric’s
> > bgp.he.net tool is now allowing users to make right atlas measurements
> > without any authentication to RIPE Atlas,  this was surprising to me
> > as when I asked a NCC staff member at the last meeting in Berlin about
> > doing something similar to this I was explicitly told that this would
> > not be desired.
> >
> > I cannot remember who told me this ( and I don't have an easy way of
> > finding out again )  but assuming this is allowed I will happily copy
> > this for bgp.tools
> >
> > But I'd like to check with Atlas people before doing such a thing.
> >