Re: [atlas] Why has probe growth stagnated?
Hi Robert, all: One suggestion is to reach out to community networks operating in Europe and beyond to help ensure their networks are being monitored as well. Let me know if you need me to facilitate any introductions to the wider community via the IGF Dynamic Coalition on Community Connectivity (DC3) email list. Best, -Michael On Fri, Mar 1, 2019 at 3:15 PM Robert Kisteleki wrote: > Hello, > > On 2019-02-20 09:36, Gert Doering wrote: > > Hi, > > > > On Wed, Feb 20, 2019 at 06:52:28AM +0200, Hank Nussbacher wrote: > >> On Tue, 19 Feb 2019, s...@gibbard.org wrote: > >> > >> ASN coverage is just 5.6%. That really doesn't give a complete global > >> view for stats. Rather than return money to LIRs every year in their > >> bill, why not state that any LIR running an active probe within their > ASN > >> will get a 50Euro credit on their bill? That alone would increase ASN > >> footprint coverage quickly within RIPE. > > > > I like that suggestion :-) > > > > (Not sure how that will pan out if we reach 50% of all LIRs, but to get > > from 5% to 20%, it might definitely be an idea) > > Interesting idea indeed! > > Let me try to address a few points that were mentioned before, perhaps > by providing some background information along the way. > > The growth rate is indeed slower than in the early days. As with > everything, there's no single reason for this. On one hand we had > shortage of hardware probes (which is hopefully solved for the moment, > see: > > https://labs.ripe.net/Members/alun_davies/new-ripe-atlas-version-4-probes/document_view_resolve > ). > On the other hand, it's harder to reach "uncovered" networks and as > someone mentioned before, the "the low hanging fruit being pretty > saturated". > > We measure ASN coverage compared to all ASN-s that announce anything at > all, so depending on how hard one wants to argue for having a probe "in > every AS", this may or may not need improvement. We're certainly > encouraging all our ambassadors and prospective hosts to deploy probes > where we don't have them yet. The current application process has a > built-in preference for new networks. > > For a few years now the procurement of new probes is funded exclusively > by sponsors. Therefore we'd be very happy to see more sponsors stepping up. > > Finally: the public graph about the number of users shows the total we > have so far (meaning users who interacted with RIPE Atlas while they > were logged in via RIPE Access). We can, if this is deemed useful, track > the number of recent users too (for some definition of recent). > > Regards, > Robert > >
Re: [atlas] Why has probe growth stagnated?
Hello, On 2019-02-20 09:36, Gert Doering wrote: > Hi, > > On Wed, Feb 20, 2019 at 06:52:28AM +0200, Hank Nussbacher wrote: >> On Tue, 19 Feb 2019, s...@gibbard.org wrote: >> >> ASN coverage is just 5.6%. That really doesn't give a complete global >> view for stats. Rather than return money to LIRs every year in their >> bill, why not state that any LIR running an active probe within their ASN >> will get a 50Euro credit on their bill? That alone would increase ASN >> footprint coverage quickly within RIPE. > > I like that suggestion :-) > > (Not sure how that will pan out if we reach 50% of all LIRs, but to get > from 5% to 20%, it might definitely be an idea) Interesting idea indeed! Let me try to address a few points that were mentioned before, perhaps by providing some background information along the way. The growth rate is indeed slower than in the early days. As with everything, there's no single reason for this. On one hand we had shortage of hardware probes (which is hopefully solved for the moment, see: https://labs.ripe.net/Members/alun_davies/new-ripe-atlas-version-4-probes/document_view_resolve). On the other hand, it's harder to reach "uncovered" networks and as someone mentioned before, the "the low hanging fruit being pretty saturated". We measure ASN coverage compared to all ASN-s that announce anything at all, so depending on how hard one wants to argue for having a probe "in every AS", this may or may not need improvement. We're certainly encouraging all our ambassadors and prospective hosts to deploy probes where we don't have them yet. The current application process has a built-in preference for new networks. For a few years now the procurement of new probes is funded exclusively by sponsors. Therefore we'd be very happy to see more sponsors stepping up. Finally: the public graph about the number of users shows the total we have so far (meaning users who interacted with RIPE Atlas while they were logged in via RIPE Access). We can, if this is deemed useful, track the number of recent users too (for some definition of recent). Regards, Robert
[atlas] Graphing packet loss
I've got some ping measurements set up and am trying to find a good way to visualise outages. The Latencymon tool obviously shows red shaded sections on the graphs for an outage, but they vanish when zoomed out a long way. A graph of average packet loss seems the obvious way of doing it, but it doesn't look like there's any way to get this from the Atlas website, sort of downloading the data and building my own tool to analyse it. Am I missing something obvious? :) -- - Steve
